December 2020 - IJG

Miettinen, J. and Liew, S. C., 2009, Estimation of Benchmark Map of Forest Carbon Stocks in Biomass Distribution in Peninsular Malaysia and Tropical Regions Across Three Continents. in the Islands of Sumatra, Java and Borneo Proceedings of the National Academy of Based on Multi-Resolution Remote Sensing Sciences of the United States of America, Vol. Land Cover Analysis. Mitigation and Adaptation 108(24), 9899-9904. DOI 10.1073/pnas.1019- Strategies for Global Change, Vol. 14(4), 357- 576108. 373. doi: DOI 10.1007/s11027-009-9169-6. Sader, S. A., Waide, R. B., Lawrence, W. T. and Joyce, A. T., 1989, Tropical Forest Biomass and Montesano, P. M., Rosette, J., Sun, G., North, P., Successional Age Class Relationships to a Nelson, R. F., Dubayah, R. O, Ranson, K. J. and Vegetation Index Derived from Landsat Tm Kharuk, V., 2015, The Uncertainty of Biomass Data. Remote Sensing of Environment, Vol. 28, Estimates from Modeled ICESat-2 Returns 159-198. Across a Boreal Forest Gradient. Remote Scaramuzza, P., Micijevic, E. and Chander, G., Sensing of Environment, Vol. 158, 95-109. doi: 2009, SLC Gap-Filled Products Phase One DOI 10.1016/j.rse.2014.10.029. Methodology 20042009. Available from http://landsat.usgs.gov/. Moser, G., Leuschner, C., Hertel, D., Graefe, S., Schultz, P. A. and Halpert, M. S., 1995, Global Soethe, N. and Iost, S., 2011, Elevation Effects Analysis of the Relationships among a Previous on the Carbon Budget of Tropical Mountain Termvegetation Index, Next Term Precipitation, Forests (S Ecuador): The Role of the and Land Surface Temperature. International Belowground Compartment. Global Change Journal of Remote Sensing, Vol. 16(15), Biology, Vol. 17(6), 2211-2226. doi: DOI https://doi.org/10.1080/01431169508954590. 10.1111/j.1365-2486.2010.02367.x. Smith, W. B., 1986, Factors and Equations to Estimate Forest Biomass in the North-Central Okuda, T., Suzuki, M., Numata, S., Yoshida, K., Region. USDA Forest Service North Central Nishimura, S., Adachi, N., Niiyama, K., Forest Experiment Station Research Paper, Vol. Manokaran, N. and Hashim, M., 2004, 268,1-6. Estimation of Aboveground Biomass in Logged Solberg, S., Næsset, E., Gobakken, T. and and Primary Lowland Rainforests Using 3-D Bollandsås, O. M., 2014, Forest Biomass Photogrammetric Analysis. Forest Ecology and Change Estimated from Height Change in Management, Vol. 203(1-3), 63-75. DOI Interferometric SAR Height Models. Carbon 10.1016/j.foreco.2004.07.056. Balance and Management, Vol. 9(1), 5. doi: 10.1186/s13021-014-0005-2. Qi, J., Chehbouni, A., Huete, A. R., Kerr, Y. H. and Spanner, M. A., Pierce, L. L., Peterson, D. L. and Sorooshian, S., 1994, A Modified Soil Adjusted Running, S. W., 1990, Remote-Sensing of Vegetation Index. Remote Sensing of Temperate Coniferous Forest Leaf-Area Index- Environment, Vol. 48(2), 119-126. the Influence of Canopy Closure, Understory Vegetation and Background Reflectance. Rai, S. N. and Proctor, J., 1986, Ecological Studies International Journal of Remote Sensing, Vol. on Four Rainforests in Karnataka, India: II. 11(1), 95-111. Litterfall. Journal of Ecology, Vol. 74(2), 455- Stape, J. L., Binkley, D. and Ryan, M. G., 2008, 463. Production and Carbon Allocation in a Clonal Eucalyptus Plantation with Water And Nutrient Riegel, J. B., Bernhardt, E. and Swenson, J., 2013, Manipulations. Forest Ecology and Estimating Above-Ground Carbon Biomass in a Management, Vol. 255(3-4), 920-930. doi: DOI Newly Restored Coastal Plain Wetland Using 10.1016/j.foreco.2007.09.085. Remote Sensing. Plos One, Vol. 8(6). DOI Steininger, M. K., 2000, Satellite Estimation of 10.1371/journal.pone.0068251. Tropical Secondary Forest Above-Ground Biomass: Data from Brazil and Bolivia. Rouse, J. W., Hass, R. H., Schell, J. A. and Deering, International Journal of Remote Sensing, Vol. D. W., 1973, Monitoring Vegetation Systems in 21(6-7), 1139-1157. the Great Plains with ERTS. Paper Read at Steven, M. D., Malthus, T. J., Baret, F., Xu, H. and Proceedings of the third ERTS Symposium. Chopping, M. J., 2003, Intercalibration of NASA SP-351, 301–317. Vegetation Indices from Different Sensor Systems. Remote Sensing of Environment, Vol. Roy, P. S. and Ravan, S. A., 1996, Biomass 88(4), 412-422. DOI 10.1016/j.rse.2003.08.010. Estimation Using Satellite Remote Sensing Data-An Investigation On Possible Approaches For Natural Forest. Journal of Biosciences, Vol. 21(4), 535-561. Saatchi, S. S., Harris, N. L., Brown, S., Lefsky, M., Mitchard, E. T. A., Salas, W., Zutta, B. R., Buermann, W., Lewis, S. L., Hagen, S., Petrova, S., White, L., Silman, M. and Morel, A., 2011, International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Thenkabail, P. S., Enclona, E. A., Ashton, M. S., Yemefack, M., Bijker, W. and De Jong, S. M., Legg, C. and De Dieu, M. J., 2004, Hyperion, 2006, Investigating Relationships between IKONOS, ALI and ETM Plus Sensors in the Landsat-7 ETM+ Data and Spatial Segregation Study of African Rainforests. Remote Sensing of of LULC Types Under Shifting Agriculture in Environment, Vol. 90(1), 23-43. doi: DOI Southern Cameroon. International Journal of 10.1016/j.rse.2003.11.018. Applied Earth Observation and Geoinformation, Vol. 8(2), 96-112. DOI 10.1016/j.jag- Thompson, S. K.,1990, Adaptive Cluster Sampling. .2005.08.003. Journal of the American Statistical Association, Vol. 85 (412):1050-1059. Zhang, J. and Fu, C., 1999, A Study on Relationships between Remote Sensing Tian, X., Su, Z. B., Chen, E. X., Li, Z. Y., van der Information and Plant Photo-Synthetic Tol, C., Guo, J. P. and He, Q. S., 2012, Parameters in Estimating Biomass Model. Acta Estimation of Forest Above-Ground Biomass Geodaetica et Cartographica Sinica, Vol. 28(5). Using Multi-Parameter Remote Sensing Data Over a Cold and Arid Area. International Journal Zhang, J., Ge, Y., Chang, J., Jiang, B., Jiang, H., of Applied Earth Observation and Peng, C. H., Zhu, J. R., Yuan, W. G., Qi, L. Z. Geoinformation, Vol. 14(1), 160-168. DOI and Yu, S. Q., 2007, Carbon Storage by 10.1016/j.jag.2011.09.010. Ecological Service Forests in Zhejiang Province, Subtropical China. Forest Ecology and Tsutsumi, T., Yoda, K., Sahunalu, P., Management, Vol. 245(1-3), 64-75. DOI Dhanmanonda, P. and Prachaiyo, B., 1983, 10.1016/j.foreco.2007.03.042. Forest: Felling, Burning and Regeneration. In Shifting Cultivation, 13-62. Tokyo. Zhao, J. F., Yan, X. D. and Jia, G. S., 2012, Simulating Net Carbon Budget of Forest Viriyabuncha, C., Vacharangkura, T., Petmak, P. Ecosystems and its Response to Climate Change and Ishizuka, M., 2002, The Evaluation System in Northeastern China Using Improved for Carbon Storage in Forest Ecosystem in FORCCHN. Chinese Geographical Science, Thailand (I. above ground biomass). edited by Vol. 22(1), 29-41. DOI: 10.1007/s11769-012- Sivilculture Research Division: Royal Forest 0512-6. Department. Zheng, D. L., Rademacher, J., Chen, J. Q., Crow, T., Xu, M., Cao, C. X., Tong, Q. X., Li, Z. Y., Zhang, Bresee, M., le Moine, J. and Ryu, S. R., 2004, H., He, Q. S., Gao, M. X., Zhao, J. A., Zheng, Estimating Aboveground Biomass Using S., Chen, W. and Zheng, L. F., 2010, Remote Landsat 7 ETM+ Data Across a Managed Sensing Based Shrub Above-Ground Biomass Landscape in Northern Wisconsin, USA. and Carbon Storage Mapping in Mu US Desert, Remote Sensing of Environment, Vol. 93(3), China. Science China-Technological Sciences, 402-411. DOI 10.1016/j.rse.2004.08.008. Vol. 53,176-183. DOI 10.1007/s11431-010- 3231-1. Zheng, X. Q., Shen, F. J., Zheng, S. N., Argaman, E., Blumberg, D., Ben-Asher, J. and Amir, S., 2012, Estimation of Regional Evapotranspiration and Biomass Production from Remote Sensing Data by Artificial Neural Network (ANN) Method. Journal of Food Agriculture & Environment, Vol 10(3-4), 1558-1561. International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Improving the Accuracy of GNSS Data in the Absolute Point Positioning Based on Linear Relational Model Elshewy M. A.,1 Hamdy A. M.,1 Elsheshtawy A. M.1 and Yunusov A. G.2 1Department of Civil Engineer, Faculty of Engineering, Al-Azhar University, Cairo, Egypt E-mail: [email protected], [email protected], [email protected] 2Department of Geodesy and Geoinformatics, Faculty of Urban Cadaster, State University of Land Use Planning, Moscow, Russia, E-mail: [email protected] Abstract The aim of this study is to improve GNSS based absolute point positioning accuracy to adequate many applications and reduce the observation time of surveying depending on the Improved Absolute Positioning (IAP) technique. IAP it's a new technique that uses the Linear Relational (LR) model or any other model to improve the accuracy of GNSS data in absolute positioning. Performance evaluation of the IAP technique was based on a compilation of different statistical parameters and goodness-of-fit measures and then outcomes of the LR model in different cases were compared. This study showed a series of significant improvements in the GNSS data in absolute positioning based on the IAP technique. Therefore, it can be applied for improving the accuracy of coordinates from the absolute point positioning in various works and presents suitable accuracy for different engineering applications. In addition, this technique introduces a low-cost technique as a result of reducing the observation time and using a single receiver. 1. Introduction logistical capacity and flexibility processing the The Global Navigation Satellite System (GNSS) is absolute positioning determination as well as used widely in various applications, from low- providing accurate position solutions from resolution navigation to high-accuracy positioning centimeters to the decimeter level (Constantin, (Gülal et al., 2015). GNSS has many features 2011). compared to conventional survey techniques (Xu, 2012). According to the measurements made on the Lately, many types of research have been tended GNSS signals, the receiver positioning can be to develop the absolute positioning technique, classified into two techniques, single point because it is a low-cost measurement technique. positioning (SPP) and relative positioning (RP) One of these developments is Precise Point (Satirapod and Kriengkraiwasin, 2006). When Positioning (PPP). In the PPP approach, GNSS navigation observations are performed in one observations from a single receiver are used to particular station, positioning technique is referred estimate the receiver location, ambiguity, receiver to as single point positioning. Instead of the term clock shift and wet tropospheric delay (Zumberge et \"single point position\", we can use the term al., 1997 and Kouba and Héroux, 2001). Several \"absolute point position\" (Hofmann et al., 2001). studies have examined attempts to improve PPP The relative positioning technique requires tracking accuracy, PPP requires at least 20-minute to achieve the same satellites simultaneously using two 10-centimeter positioning accuracy (Fang et al., receivers, one as a reference station and the other as 2001 and Gao and Chen, 2004). Satirapod and a Rover station, in order to find Rover coordinates Homniam (2006), recently developed simple PPP with respect to the reference station (El-Rabbany, technique provides accuracy lower than one meter, 2002). So called differential Global Positioning by 15-minute observation period using a dual- System (DGPS), this technique is applied to high frequency GNSS receiver (Satirapod and Homniam, accuracy positioning measurement. However, the 2006). Static PPP takes 60-minute to reach 5- effectiveness of the differential positioning centimeter horizontal accuracy at a 95% confidence technique generally depends on the distance level and 24 hours to achieve 1-centimeter between the two receivers. Also, the cost of horizontal accuracy at a 95% confidence level differential positioning technique is still much more (Abou-Galala et al., 2018). PPP needs expensive than the absolute point positioning. improvements to reduce the very long initial Therefore, from a practical point of view, there convergence period (Abou-Galala et al., 2018). It is should be a measurement technique that contains noted from previous studies that PPP is important International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

and used in many applications, but one of its major 2.3.1 Linear Relation (LR) model drawbacks is the need for a long initial convergence LR model consists of two-step, in the first step, the period to obtain appropriate accuracy. mean values (CX, CY, and Ch) of the differences between the coordinates of the control points from Therefore, the main objective of this study is to Total Station data (XT, YT, and hT) and from the improve GNSS based absolute point positioning GNSS data in absolute positioning (XG, YG, and hG) accuracy and allows users to obtain reliable were determined from equations 1, 2 and 3: coordinates with accuracies expressed in decimeters derived from rapid data acquisition as a few seconds ������������ = ������������������ depending on IAP technique. ������ 2. Materials and Methods Equation 1 2.1 Study Area and Measurements The study area is located in Madinaty city, Cairo, where VX = XG – XT Egypt, with an area of about 150 m by 100 m. The observations were carried out on February 02, 2020. ������������ = ������������������ Two static GNSS measured points (CP1 and CP2) ������ were used as an occupied and back sight points to tie the Total Station coordinates to the WGS 84 Equation 3 system/UTM zone 36 North as map projection, with EGM-2008 geoid model, and their coordinates were where VY = YG – YT (368127.482, 3327720.167, 280.873) and (368089.773, 3327762.849, 278.498). The total time ������ℎ = ������������ℎ of the GNSS measurements is approximately 3 ������ hours where each point was occupied by the receiver of about 20 seconds. The area was divided Equtation into 165 points with a distance of about 10 m between every two consecutive points. Figure 1 where Vh = hG – hT illustrates the study area plan. The coordinates of these 165 points were determined by following In the second step, the mean values (CX, CY, and Ch) instruments: which determined from equations 1, 2 and 3 were used to generate the improved coordinates (XI, YI, • A Sokkia CX-105 Total Station was used and hI) for the rest points by using equations 4, 5 to measure the coordinates of points. and 6: • To collect coordinates from GNSS data, XI = XG - CX Equation 4 SOKKIA Grx2 GNSS receiver was YI = YG – CY Eqaution 5 adopted. hI = hG - Ch Equation 6 2.2 Research Methodology 2.4 Cases of Select Points for Obtained the Relation The methodology of enhancing the accuracy of The testing was carried out using 165 points and GNSS data in the absolute positioning based on the their coordinates were determined from Total IAP technique includes data acquisition task, Station and GNSS data; These points divided into determination of correlation between GNSS data two groups; the first group (coordinates of some and Total Station for control points, create the LR points) used to find the relation between Total model, improve the rest of GNSS data by using the Station and GNSS data. Figure 3 shows the numbers LR model. These steps are further elaborated in the and distribution of the 165 points. The second group following sub-sections. Figure 2 shows the detection used to evaluate the results. Four different cases flow chart of improving the accuracy of GNSS data were used to select the first group which used to using absolute positioning based on the IAP find the relation between Total Station and GNSS technique. data: 2.3 Improved Absolute Positioning (IAP) Technique • Case 1, the first 5 points, about 3% of all The IAP it's a new technique that uses the LR model the data. or any other model to improve the accuracy of GNSS data in absolute positioning. In this • Case 2, the first 10 points, about 6% of all technique, the LR model was created from of 3-5 the data. control points obtained from Total Station data and GNSS data in absolute positioning. • Case 3, the first 20 points, about 12% of all the data. • Case 4, the first 30 points, about 18% of all the data. International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Figure 1: Map of the study area Figure 2: The detection flow chart for improving the accuracy of GNSS data in absolute positioning based on the IAP technique Figure 3: Distribution of the studied points International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

3. Results and Discussion 3.2 Improving the Accuracy of GNSS Data Using Firstly, the coordinates of all the points from Total the IAP Technique Station and GNSS data were determined and the Firstly, the mean values (CX, CY, and Ch) of the differences between them were derived. Secondly, differences between Total Station data and GNSS in the relation between the Total Station and GNSS absolute point positioning technique data by using data for control points was expressed using the LR equations 1, 2, and 3 for control points were model. After that, this relation was used to generate determined for the four cases. See Table 2. the improved coordinates of the rest of the GNSS data. Finally, to evaluate the results, the improved After that, the values of CX, CY, and Ch in the GNSS coordinates were compared with the Total four cases were used to generate the improved Station coordinates. coordinates of the rest GNSS in absolute point positioning technique data by using equations 4, 5, 3.1 Accuracy of GNSS Data of All the Points and 6. To evaluate the accuracy of the results; the At this step, the mean value, absolute mean value, descriptive statistics of differences between and RMSE of the differences between Total Station coordinates of Total Station and improved and GNSS data in absolute point positioning coordinates of GNSS data were determined and technique of all points were computed and tabulated summarized in Tables 3, 4, and 5. The results in the in Table 1. The results of the table show that the tables showed that, when using the LR model to average difference in position between Total Station express the relation between Total Station data and and GNSS data is about 4.059m. The accuracy of GNSS in absolute positioning data and generate new GNSS data in absolute point positioning with an coordinates from the rest of GNSS data, the mean observation duration of 20 seconds is 4.069m which differences, the absolute mean differences, and does not match the required accuracy of the geodetic RMSE improved in the four cases. There is no works. Therefore, the important question here, is it significant difference between the improvements in possible to improve the accuracy of GNSS data in the four cases in the X-direction and the Y- absolute point positioning technique with a duration direction. of observation of about 20 seconds?. Table 1: Descriptive statistics of differences between Total Station and GNSS data of all the points X-direction Mean value Absolute mean value RMSE (m) Y-direction h-direction of differences (m) of differences (m) 3.864 3.861 3.861 0.386 Position 0.351 0.351 1.216 -1.066 1.081 4.069 4.059 4.059 Table 2: Value of the mean values (CX, CY, and Ch) in the four cases CX (m) case 1 case 2 case 3 case 4 CY (m) Ch (m) 3.948 3.875 3.841 3.824 0.227 0.163 0.164 0.172 -1.757 -1.435 -1.248 -1.205 Table 3: Descriptive statistics of differences between original/improved coordinates in X-direction Mean value of differences Absolute mean value of RMSE (m) (m) difference (m) case 1 original improved original improved original improved case 2 GNSS data GNSS data GNSS data GNSS data GNSS data GNSS data case 3 case 4 3.858 -0.090 3.858 0.162 3.861 0.179 3.860 -0.015 3.860 0.135 3.863 0.158 3.864 0.023 3.864 0.132 3.867 0.163 3.869 0.045 3.869 0.138 3.873 0.172 International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Table 4: Descriptive statistics of differences between original/improved coordinates in Y-direction Mean value of differences (m) Absolute mean value of RMSE (m) difference (m) case 1 original improved original improved original improved case 2 GNSS data GNSS data GNSS data GNSS data GNSS data GNSS data case 3 case 4 0.355 0.128 0.355 0.161 0.390 0.206 0.364 0.201 0.364 0.205 0.396 0.255 0.377 0.213 0.377 0.217 0.407 0.262 0.391 0.219 0.391 0.224 0.419 0.265 Table 5: Descriptive statistics of differences between original/improved coordinates in h-direction Mean value of differences (m) Absolute mean value of RMSE (m) difference (m) case 1 original improved original improved original improved case 2 GNSS data GNSS data GNSS data GNSS data GNSS data GNSS data case 3 case 4 -1.044 0.713 1.060 0.830 1.194 0.918 -1.042 0.393 1.058 0.614 1.197 0.707 -1.041 0.207 1.058 0.521 1.205 0.642 -1.035 0.170 1.035 0.529 1.211 0.652 In the h-direction, it is noted that, the more points Where: I - Statistics (mean value of differences, used to create the LR model, the greater accuracy is absolute mean value of differences and RMSE) of obtained. Additionally, it is not necessary for improved coordinates from LR model. creating the LR model as much of the control points, but only about 3-5 control points suffice as in case O - Statistics (mean value of differences, 1. As mentioned by Abdallah and Schwieger (2014) absolute mean value of differences and RMSE) of they obtained the RMSE in 3 dimensions of 10 cm original coordinates from rest of GNSS in absolute from 10 minutes initialization time by using GIPSY- positioning data. The results of the previous figures OASIS software, as well as CSRS-PPP online showed that when using the IAP technique to service (Abdallah and Schwieger, 2014). By improve the GNSS in absolute positioning data; the comparing our study results with these results, here mean value of differences in X, Y, and h directions a lower accuracy was obtained, but the observing was improved with a percentage of about 99%, time was only about 20 seconds. In addition, 44%, and 75% respectively. The absolute mean according to the results obtained by Kim and Park, value of differences was improved with a the horizontal and vertical RMSE along the test percentage of about 97%, 43%, and 48% route were 0.53 and 0.69 m, respectively, to respectively. While RMSE was improved with a examine their PPP algorithm for a moving platform percentage of about 96%, 36%, and 45% (Kim and Park, 2017). Also, by comparing our respectively. In general, the percentages of study results with these results, horizontal accuracy improvement in the four cases about 59%, 63%, here increased to 0.208m while the vertical 66%, and 66% respectively. It is noted that there is resolution was decreased to 0.730m. no significant difference in the percentage of improvement between the four cases in X-direction 3.3 Percentage of Improvement of GNSS Data with and Y-direction, but in h-direction, the results in the IAP Technique cases 2, 3 and 4 improved twice as much as in the The percentages of improvement in the accuracy of first case. In addition, when using the IAP technique coordinates of the points from GNSS in absolute to improve the GNSS in absolute positioning data; positioning data using the LR model have been the mean value of differences in position improved calculated using equation 7, the results illustrate in from 4.059 m to 0.656 m by a percentage of about Figures 4, 5, and 6: 84%. While the RMSE improved from 4.069 m to 0.761 m by a percentage of about 81%. Improvement percentage = [ 1 – (I / O)] % Equation 7 International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Figure 5: Percentages of improvement in the absolute mean values of GNSS in absolute positioning data using the IAP technique Figure 6: Percentages of improvement in the RMSE of GNSS in absolute positioning data using the IAP technique 4. Conclusions GNSS data. To improve the accuracy of GNSS data In recent years, the GNSS has been applied in in absolute positioning, the IAP technique has been various aspects of life and it becomes an influencing used. IAP is a new technique that uses the LR model factor in many engineering applications. The cost of or any other model to improve the accuracy of differential positioning technique is still much more GNSS data in absolute positioning. The LR model expensive than the absolute point positioning. was used to express this relation and to achieve the Therefore, many studies have been attempted to required improvement. The results show that the improve the accuracy of GNSS in the absolute point accuracy of GNSS data improved with the IAP positioning technique. The purpose of this paper is technique in X, Y, and h directions by about 96%, to improve the accuracy of GNSS data in absolute 36%, and 45% respectively. By using the IAP positioning and make it convenient for various technique, the position of points improved by about applications during a rapid survey session of a few 81%. In order to create the LR model It's not seconds. The study idea depends on finding the necessary using a lot of control points, but enough relation between Total Station and GNSS data in the about 3-5 points. It is well known that the correction absolute positioning for some control points (3-5 factors differ according to the observing times and points) and use this relation to improve the rest of location, therefore it is preferable to carry out new International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

correction factors in case of using different Gülal, E., Dindar, A. A., Akpınar, B., Tiryakioğlu, observing times or other study areas. Further İ., Aykut, N. O. and Erdoğan, H., 2015, Analysis research will be conducted on larger areas, varying and Management of GNSS Reference Station environments and conditions to give a better Data. Tehnički Vjesnik, Vol. 22(2), 407-414. assessment of the application of the IAP technique. Hofmann, B., Lichtenegger, H. and Collins, J., References 2001, GPS Theory and Practice. Springer Wien NewYork. Abdallah, A. and Schwieger, V., 2014, March. Accuracy Assessment Study of GNSS Precise Kim, M. and Park, K. D., 2017, Development and Point Positioning for Kinematic Positioning. In Positioning Accuracy Assessment of Single- Schattenberg, J., Minßen, TF: Proceeding on 4th Frequency Precise Point Positioning Algorithms International Conference on Machine Control by Combining GPS Code-Pseudorange and Guidance (MCG), Braunschweig: Institut Measurements with Real-Time SSR Corrections. für mobile Maschinen und Nutzfahrzeuge, Sensors, Vol. 17(6), 1347. Braunschweig, Germany. Kouba, J. and Héroux, P., 2001, Precise Point Abou-Galala, M., Rabah, M., Kaloop, M. and Zidan, Positioning Using IGS Orbit and Clock Z. M., 2018, Assessment of the Accuracy and Products. GPS Solutions, Vol. 5(2), 12-28. Convergence Period of Precise Point Positioning. Alexandria Engineering Journal, Satirapod, C. and Kriengkraiwasin, S., 2006, Vol. 57(3), 1721-1726. Performance of Open Source Precise Point Positioning Software Using Single-Frequency Constantin, O. A., 2011, Precise Point Positioning GPS Data. Artificial Satellites, Vol. 41(2), 79- Applicability for the Implementation of a 86. Cadastral System in IASI Municipality. Published in Proceedings of the GeoPreVi 2011 Satirapod, C. and Homniam, P., 2006, GPS Precise Conference, 65-78, May 12-13, 2011, Bucharest, Point Positioning Software for Ground Control Romania. Point Establishment in Remote Sensing Applications. Journal of Surveying Engineering, El-Rabbany, A., 2002, Introduction to GPS: the Vo. 132(1), 11-14. Global Positioning System. Artech House. Xu, H., 2012, Application of GPS-RTK Technology Fang, P., Gendt, G., Springer, T. and Mannucci, T., in the Land Change Survey. Procedia 2001, IGS Near Real-Time Products and their Engineering, Vol. 29, 3454-3459. Applications. GPS Solutions, Vol. 4 (4), 2-8. Zumberge, J. F., Heflin, M. B., Jefferson, D. C., Gao, Y. and Chen, K., 2004, Performance Analysis Watkins, M. M. and Webb, F. H., 1997, Precise of Precise Point Positioning Using Real-Time Point Positioning for the Efficient and Robust Orbit and Clock Products. Journal of Global Analysis of GPS Data from Large Networks. Positioning Systems, Vol. 3(1-2), 95-100. Journal of Geophysical Research: Solid Earth, Vol. 102(B3), 5005-5017. International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Comparison and Validation of VTEC Derived from GPS, IRI-Plas and NeQuick-2 During 2015 and 2019 in India Alluri, V.,1 Chowdhary. V. R.,1 Joshi, S.,1,2 Banerjee, S.3 and Kothari, N.1 1Department of Electronics and Telecommunication, International Institute of Information Technology Pune, India 2Department of Computer Science, Savitribai Phule Pune University, Pune, India 3Department of Computer Engineering, International Institute of Information Technology, Pune, India Abstract The paper compares Vertical Total Electron Content (VTEC) values extracted from IGS’s distributed network of dual-frequency GPS stations in India, and the VTEC values extracted through IRI-Plas-2017 and NeQuick- 2.0.2 models at the same locations and the same durations of time. Diurnal variation and relative deviation of modelled and measured VTEC of IGS reference stations at Lucknow (LCK3, Latitude: 26.91218, Longitude: 80.95564) and Hyderabad (HYDE, Latitude: 17.41728, Longitude: 78.55088) are visualized through graphical representation, during 2015 and 2019. Further, statistical analysis was performed on both datasets. The outputs of this project revealed that, the magnitude of maximum relative deviation of modelled VTEC from measured VTEC was high while using IRI-Plas model at Lucknow and Hyderabad in each month of solar maximum and solar minimum. Furthermore, during solar minimum, VTEC is highly overestimated by both models during peak hours of ionization and the magnitude of overestimation at these hours is higher while using IRI-Plas model for both regions. Finally, the highest coefficient of determination value was recorded at Hyderabad during 2019 while using IRI-plas model. The R2 analysis shows that IRI-Plas model produces a more accurate representation of VTEC during solar minimum and maximum at both regions. 1. Introduction modelling techniques and newer datasets, updated Total Electron Content (TEC) has been an important versions of IRI such as: IRI-1985, IRI-1990, IRI- parameter to study disturbances in the ionosphere 2000, IRI-2007, IRI-2016 and most recently IRI- that are caused by variations in intensity of solar Plas, were launched (Bilitza, 1990, Bilitza et al., radiation. While comprehensive analysis on TEC 2000 and Bilitza and Reinisch, 2008). TEC, is one derived from trans-ionospheric radar instruments amongst the thirty-seven other ionospheric started as early as 1957 (Evans, 1957 and Bauer and parameters that are calculated by IRI. Similarly, Daniels, 1959), with the advent of artificial NeQuick is another prominent ionospheric model satellites, TEC derived from Global Navigation that is developed by the International Centre for Satellite Systems (GNSS) is currently the most Theoretical Physics (ICTP) in Italy, along with the preferred data source for observing ionospheric University of Graz in Austria. It is a three- behaviour (Mendillo, 2005). Over the last 20 years, dimensional and time-dependent empirical model of while several researchers offered successful insight the ionosphere’s electron density profile. Unlike the on the diurnal, monthly, seasonal and annual IRI-model, TEC and electron density are the only variations in TEC at various latitudinal regions two parameters calculated by NeQuick. The latest (Bhattacharya et al., 2009, Unnikrishnan et al., 2002 version of this model is NeQuick-2. and D’ujanga et al., 2012), several others offered an insight into the electromagnetic phenomena In this study, Vertical Total Electron Content responsible for short-term and long-term impacts of (VTEC) extracted from GPS, IRI-Plas and solar radiation on ionospheric TEC (Chauhan and NeQuick-2 are compared at two regions in India Singh, 2010, Anderson et al., 2006 and Appleton, during 2015 and 2019. The comparative and 1946). correlation analysis aims to quantify deviation of modelled TEC data from measured TEC data, at a Based on years of research on ionospheric mid-latitude region and a low-latitude region of plasma, a data-based empirical model named India, during the solar maximum (2015) and solar International Reference Ionosphere (IRI) was first minimum (2019). launched in 1978, as a standard for ionospheric parameters. Over time, with the emergence of better International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

2. Literature Aeronautics and Space Administration (NASA). 2.1 Total Electron Content (TEC) IGS station in Lucknow (LCK3) is located at TEC is a parameter widely used to depict the effect 26.91218 N and 80.95564 E, and IGS station in of solar radiation on the ionosphere. TEC can be Hyderabad (HYDE) is located at 17.41728 N and defined as the total number of electrons that are 78.55088 E. present in an area enclosed by a tube with an arbitrary cross-section of 1 square meter laid over 2.4 TEC Derived from Dual-Frequency GPS the entire length between the satellite and receiver (Okoh et al., 2015). TEC values will be high during In a GNSS network such as the GPS, for the the daytime due to ionization by X-rays and UV- ‘navigation message’ to travel from the satellite to rays, and low during night-time due to the the receiver, a ‘carrier wave’ is required. In case of a recombination process. Further, when Interplanetary Magnetic Fields (IMF) directed towards earth dual-frequency GPS design, two carrier waves are interact with the magnetosphere, electromagnetic processes govern TEC enhancements and depletions used: L1 at 1575.42 MHz and L2 at 1227.60 MHz. in the ionosphere. Also, TEC is directly proportional Pseudo-range estimations from ‘L1’ and ‘L2’ are to signal delay because the increase in free electrons in the ionosphere creates a highly dispersive used to calculate Slant TEC (STEC) at a station. The medium through which GPS signals must travel. empirical formula for calculating STEC from 2.2 Solar Cycle The intensity of solar activity increases and pseudo-range measurements is given below: decreases in an 11-year cycle. This solar cycle dictates the extremity of X-rays and UV-rays ������������������������ = 2 [������1���2���1 2������22 2 ] (������2 − ������1) + ������������ + ������ ������ emitted by the sun. Therefore, the solar cycle has ������ − ������2 dramatic implications on the electromagnetic mechanisms of the Earth’s upper atmosphere Equation 1 (David, 2015). In a typical solar cycle, intensity of solar activity increases through the first five or six Where, ‘k’ is a constant whose value is 80.62 years until it reaches a maximum, and then (m3/s2); ‘f1’ and ‘f2’ are frequencies of pseudo- decreases through the remainder of the 11-year ranges ‘P1’ and ‘P1’; ‘τr’ and ‘τs’ are the differential cycle until it reaches a minimum. The current solar cycle i.e. solar cycle 24 is predicted to end in 2020. code bias and inter-frequency bias corresponding to The cycle’s solar maximum was reached in 2014- ‘P1’ and ‘P2’ (Kenpankho et al., 2011). Once STEC 15. is corrected for the ‘satellite bias and ‘receiver bias’, 2.3 International GNSS Service (IGS) VTEC is calculated using the following formula: IGS provides highly precise navigation information through over 400 global permanent GNSS stations ������������������������ = ������������������������ {������������������ [������������������������������������ (������������ ������������ ℎ������ ������������������������)]} and close to 200 organizations spread over 100 + countries are responsible for contributing towards the establishment of this organization (source: Equation 2 www.igs.org). The accuracy and precision of GNSS measurements is very high. GNSS data derived from Where, χ is the zenith angle at receiver position, Re the IGS network fundamentally measures two is the mean radius of the earth and hm is the height atmospheric parameters: The Tropospheric Zenith of ionospheric layer (Tariku, 2015). Path Delay (ZPD) and Ionospheric TEC. This is done by combining pseudo-range measurements of 2.5 International Reference Ionosphere (IRI) GNSS with IGS precise clocks and orbits (Kouba, IRI is an empirical model developed by the 2009). IGS hosts data obtained predominantly from Committee on Space Research (COSPAR) and a single satellite navigation system i.e., Global International Union of Radio Science (URSI) to Positioning System (GPS). More recently, data from provide standardized measurements of ionospheric Russia’s GLObal NAvigation Satellite System parameters. A team of 60+ ionospheric experts from (GLONASS) has been incorporated into the IGS different parts of the world are responsible for workflow. The data is available for download from generating the model, and introducing corrections or the IGS data portal, which is hosted by National modifications for enhancing model accuracy. The inputs for this model are provided from a variety of instruments such as: a worldwide network of ionosondes, incoherent scatter radars, topside sounder satellites, and in-situ satellite measurements (Dieter et al., 2011). IRI was standardized by International Standardization Organization (ISO) in 2014. The model is developed based on experimental evidence rather than our evolving theoretical understanding of ionospheric plasma and its behaviour. Theoretical observations are only International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

used to bridge whatever gaps are encountered in the files from IGS data portal was automated using development of this model. Therefore, if certain Python script. Navigation message files, or ‘.n’ files, geographical areas and time periods do not have an which contain ephemeris for all the GPS satellites, underlying database of ionospheric research, then were also downloaded from IGS data portal using a the ionospheric parameters estimated by IRI for that Python script. The hourly VTEC data from IRI-Plas spatial and temporal extent have a risk of being and NeQuick 2 models can be obtained for a mildly unreliable. For a given date, time and specific geographic location and hour. The IRI-Plas location, the IRI model estimates ion composition, data was obtained by compiling and running the electron density, VTEC, height of ionospheric IRI-Plas source code, written in FORTRAN, which layers, vertical ion drift and ion temperatures at is available at ftp://ftp.izmiran.ru/pub/izmiran/SP- various temporal resolutions. IM/ (Gulyaeva, 2020). The default values for all the parameters were used while running the IRI-Plas 2.6 Nequick-2 Model program. Similarly, the NeQuick-2 data was NeQuick is based on the DGR profiler model, which obtained by compiling and running the NeQuick-2 was originally proposed by G. Di Giovanni and S. source code, written in FORTRAN, which is M. Radicella in 1990. The model empirically available at https://t-ict4d.ictp.it/nequick2/source- reproduces electron density profile of the ionosphere code (Zhang et al., 2010). Lower endpoint value of using the sum of Epstein layers (Sandro, 2009). The 0km, higher endpoint value of 20200km, and the DGR model was designed to fulfil, to a reasonable daily solar index F10.7 values obtained from NASA extent, the basic criteria used to judge an empirical OMNIWeb Data Explorer available at model of the ionosphere’s electron density profile https://omniweb.gsfc.nasa.gov/form/dx1.html were w.r.t. height. These criteria were first defined by used as parameters for running the NeQuick-2 Dudeney and Kressman in 1986, which state that program. The hourly VTEC data from both the mathematical formulations of ionospheric models for Lucknow and Hyderabad for all days of parameters should be simpler than traditional 2015 and 2019 was thus obtained, and processed ionogram inversion techniques. In 1995, into a .csv (comma-separated values) file in our improvements to the original DGR model made by desired format. Radicella and Zhang allowed for estimation of VTEC (Sandro and Man-Lian, 1995). Further 3.2 Data Pre-Processing improvements and modifications were made to the GPS data downloaded from IGS data portal requires original model in 2001, 2005 and 2006. These further pre-processing in order to extract VTEC additions are reflected in the latest version of this values. Since each observation file represents data model i.e. NeQuick-2. The NeQuick model has been of one single day and one single station, around particularly successful in estimating electron density 1,460 observation files need to be downloaded for of ionosphere above 100 km. The model is currently Lucknow and Hyderabad during 2015 and 2019. adopted by the European satellite navigation system However, given that some data is missing, a total of (GALILEO) for ionospheric corrections of its single 1179 observation files are downloaded using python frequency GNSS operation. Further, the NeQuick script. TEC and satellite ephemeris data is model is adopted by International compressed in observation and navigation files Telecommunication Union (ITU) as a suitable respectively in a Receiver Independent Exchange model for estimating ionospheric parameters (RINEX) format. To extract this data, each RINEX (Ezquer et al., 2017). For a given date, time and observation and respective navigation file is location, the NeQuick model estimates electron provided as an input to the GPS-TEC software, density and VTEC at various temporal resolutions. designed by Dr. Gopi Seemala of the Indian Institute of Geomagnetism (https://seemala.blogsp- 3. Method of Analysis ot.com/). The output is a text file in standard text 3.1 Data Download format, which provides in a column, VTEC values GPS data was downloaded from the IGS data portal measured at one-minute intervals, for that day and hosted by NASA at: ftp://cddis.nasa.gov/gnss/data/- location. A Python program was used to read data daily/. From the data portal, ‘observation files’ with from the output files, which was processed to obtain ‘.o’ file extension can be downloaded for each day hourly averaged VTEC values for all days of 2015 of the year. This data is available since 1992 for and 2019, at Lucknow and Hyderabad. more than 300 geographic locations. To achieve the objectives of this study, ‘.o’ files from GPS stations 3.3 Data Processing located at Lucknow and Hyderabad are downloaded Once the data files with hourly-VTEC values from for all days of 2015 and 2019. Download of ‘.o’ all three data sources are ready, they are used as an International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

input to two python scripts: the first script is used to months of March and July while using the NeQuick- graphically represent diurnal variation of VTEC at 2 model. Lucknow and Hyderabad during 2015 and 2019, using all three data sources; while the second script 4.2 Diurnal Variation and Relative Deviation (D) of is used to generate a plot to represent relative Modelled and Measured VTEC at Hyderabad deviation between modelled VTEC prediction and During 2015 measured VTEC. The relative deviation (D) is From Figure 3 and Figure 4 it can be observed that, calculated using the following formula: at Hyderabad, during 2015, the relative deviation of modelled VTEC from measured VTEC is both ������ = [(������������������������������������������������ ������������������������ − ������������������������������������������������ ������������������������)] positive and negative, depending on the duration of day and month. Throughout the year, the positive ������������������������������������������������ ������������������������ relative deviation or overestimation by modelled data is observed in the post-afternoon durations of Equation 3 the day. Low and negative relative deviation is observed predominantly in the sun-lit hours of the A third Python script is used to generate a scatter day. During certain months, the negative relative plot for the measured VTEC on X-axis, and deviation or underestimation by modelled data modelled VTEC on Y-axis from IRI-Plas and extends to later UT. From Table 2 it can be NeQuick-2 in their respective plots. Linear observed that, at Hyderabad, during 2015, the regression analysis was performed on the data to maximum relative deviation of modelled VTEC determine the capability of both models in from measured VTEC was observed between 2100 accurately producing VTEC data for the specific UT and 2300 UT, or 0000 UT, throughout the year geographic locations during 2015 and 2019. while using IRI-Plas model. 4. Results Similarly, while using the NeQuick-2 model, the 4.1 Diurnal Variation and Relative Deviation (D) of maximum relative deviation was observed between Modelled and Measured VTEC at Lucknow During 2000 UT and 2300 UT, or 0000 UT, throughout the 2015 year. Finally, the maximum relative deviation from From Figure 1 and Figure 2 it can be observed that, measured VTEC for the entire year was observed in at Lucknow, the relative deviation of modelled March at 0000 UT for both models. VTEC from measured VTEC is mostly positive throughout the year at all durations of the day. This From Table 2 it can further be observed that, at indicates that, both IRI-Plas and NeQuick-2 models Hyderabad, during 2015, the maximum relative overestimate VTEC throughout the day at Lucknow deviation of modelled VTEC from measured VTEC during 2015. From Table 1 it can be observed that, during peak hours of ionization was observed in at Lucknow, during 2015, the maximum relative September (34.25%) while using IRI-Plas model; deviation of modelled VTEC from measured VTEC and in February (-22.96%) while using the was observed between 2100 UT and 2300 UT from NeQuick-2 model. Negative relative deviation from February to October while using the IRI-Plas model. measured data was observed at peak hours of Similarly, while using the NeQuick-2 model, the ionization in certain months while using both maximum relative deviation was observed between models. While negative relative deviation at peak 2100 UT and 2300 UT from February to August. hours of ionization was observed only in the months However, during the remaining months of the year, of February and March while using the IRI-Plas the NeQuick-2 model shows maximum relative model, negative relative deviation at peak hours of deviation from measured VTEC between 1300 UT ionization was observed almost throughout the year and 1600 UT. In the month of November, both while using NeQuick-2 model. models show maximum relative deviation from measured VTEC at 1300 UT. Finally, the maximum 4.3 Diurnal Variation and Relative Deviation (D) of relative deviation from measured VTEC for the Modelled and Measured VTEC at Lucknow During entire year was observed in April at 2300 UT for 2019 both models. From Figure 5 and Figure 6 it can be observed that, at Lucknow, during 2019, the relative deviation of From Table 1 it can further be observed that, at modelled VTEC from measured VTEC is both Lucknow, during 2015, the maximum relative positive and negative during most months, with the deviation of modelled VTEC from measured VTEC exception of a few months (April, May, June and during peak hours of ionization was observed in August) where positive relative deviation was September (81.94%) while using IRI-Plas model; observed at all durations of the day. and in November (55.38%) while using NeQuick-2 model. Negative relative deviation from measured data was observed at peak hours of ionization in the International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Table 1: Maximum relative deviation and relative deviation at peak hours of ionization between modelled and measured VTEC during 2015 at Lucknow Month Maximum relative Maximum relative Peak hours Relative deviation Relative deviation deviation while using deviation while using of ionization during peak hours during peak hours IRI-Plas model (%) NeQuick-2 model (%) as observed of ionization while of ionization while from using IRI-Plas using NeQuick-2 January 195.35 at 0000 UT 182.59 at 1600 UT measured model model February 187.40 at 2100 UT 101.99 at 2100 UT data March 293.32 at 2300 UT 207.36 at 2000 UT 35.41 % 49.22 % April 426.71 at 2300 UT 392.92 at 2300 UT 0700 UT 12.81 % 12.32 % May 385.35 at 2300 UT 351.19 at 2300 UT 1000 UT 9.88 % -4.62 % June 380.13 at 2200 UT 362.13 at 2200 UT 0800 UT 22.94 % 17.21 % July 319.40 at 2200 UT 210.44 at 2200 UT 0900 UT 20.90 % 11.94 % August 308.50 at 2300 UT 190.46 at 2300 UT 0900 UT 20.65 % 11.96 % September 215.46 at 2300 UT 108.59 at 1400 UT 0800 UT 17.83 % -3.21 % October 202.01 at 2100 UT 172.49 at 1400 UT 0800 UT 47.89 % 16.60 % November 180.13 at 1300 UT 177.72 at 1300 UT 0700 UT 81.94 % 48.12 % December 134.35 at 2000 UT 112.90 at 1500 UT 0900 UT 60.50 % 32.05 % 0700 UT 72.67 % 55.38 % 0600 UT 64.02 % 52.12 % 0900 UT Table 2: Maximum relative deviation and relative deviation at peak hours of ionization between modelled and measured VTEC during 2015 at Hyderabad Month Maximum relative Maximum relative Peak hours of Relative deviation Relative deviation deviation while deviation while using ionization as during peak hours during peak hours using IRI-Plas NeQuick-2 model (%) observed from of ionization while of ionization while model (%) measured data using IRI-Plas using NeQuick-2 140.36 at 2200 UT model model January 196.21 at 2300 UT 132.34 at 2200 UT 0800 UT 14.20 % 6.70 % February 188.08 at 2200 UT 846.86 at 0000 UT 1000 UT -11.44 % -22.96 % March 1239.1 at 0000 UT 681.46 at 0000 UT 1000 UT -3.66 % -17.97 % April 758.12 at 0000 UT 178.66 at 2300 UT 1000 UT 0.95 % -13.32 % May 217.41 at 2300 UT 167.84 at 2100 UT 0900 UT 20.11 % -2.73 % June 182.76 at 2100 UT 158.72 at 2200 UT 0800 UT 21.81 % -0.44 % July 261.62 at 2200 UT 88.29 at 2000 UT 0900 UT 26.11 % -7.86 % August 132.07 at 2100 UT 121.11 at 2100 UT 0900 UT 28.26 % -7.29 % September 234.94 at 2200 UT 119.87 at 2100 UT 1000 UT 34.25 % 1.82 % October 247.29 at 2300 UT 148.22 at 2100 UT 0900 UT 28.88 % -0.31 % November 243.84 at 2200 UT 166.53 at 2000 UT 0900 UT 25.16 % 2.47 % December 303.40 at 0000 UT 0900 UT 28.57 % 11.06 % Table 3: Maximum relative deviation and relative deviation at peak hours of ionization between modelled and measured VTEC during 2019 at Lucknow Month Maximum relative Maximum relative Peak hours of Relative deviation Relative deviation deviation while deviation while using ionization as during peak hours during peak hours January using IRI-Plas NeQuick-2 model observed from of ionization while of ionization while February model (%) (%) measured data using IRI-Plas using NeQuick-2 March model model April 161.05 at 0600 UT 114.98 at 0600 UT 0800 UT 122.29 % 82.41 % May 175.29 at 0700 UT 119.35 at 0700 UT 0900 UT 108.57 % 72.74 % June 144.63 at 0600 UT 112.49 at 0600 UT 0800 UT 131.81 % 100.95 % July 780.58 at 0000 UT 492.83 at 0000 UT 0800 UT 119.25 % 91.40 % August 135.37 at 1000 UT 116.39 at 1000 UT 0900 UT 123.36 % 97 % September 159.83 at 2100 UT 93.28 at 1900 UT 0700 UT 83.07 % 44.65 % October N.A. N.A. N.A. N.A. N.A. November 135.97 at 0700 UT 91.13 at 0900 UT 0800 UT 129.53 % 82.76 % December 145.25 at 0300 UT 92.57 at 0300 UT 0700 UT 91.93 % 54.45 % 174.80 at 0200 UT 128.13 at 1200 UT 0800 UT 95.85 % 72.10 % N.A. N.A. N.A. N.A. N.A. 193.52 at 0700 UT 142.48 at 0700 UT 0900 UT 130.38 % 95.16 % International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Figure 1: Diurnal variation of modelled and measured Figure 3: Diurnal variation of modelled and measured VTEC at Lucknow during 2015. The black, green and VTEC at Hyderabad during 2015. The black, green and red solid lines denote VTEC derived from GPS, IRI-Plas red solid lines denote VTEC derived from GPS, IRI- and NeQuick-2 respectively. The vertical black dashed Plas and NeQuick-2 respectively. The vertical black line denotes the peak hour of ionization as observed from dashed line denotes the peak hour of ionization as measured data observed from measured data. Figure 2: Relative deviation of modelled and measured VTEC at Lucknow during 2015. The green and red solid lines denote relative deviation from the measured data while using IRI-plas and NeQuick-2 respectively. The vertical black dashed line denotes the peak hour of ionization as observed from measured data International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Figure 4: Relative deviation of modelled and measured VTEC at Figure 5: Diurnal variation of modelled and measured VTEC at Hyderabad during 2015. The green and red solid lines denote Lucknow during 2019. The black, green and red solid lines relative deviation from the measured data while using IRI-plas denote VTEC derived from GPS, IRI-Plas and NeQuick-2 and NeQuick-2 respectively. The vertical black dashed line respectively. The vertical black dashed line denotes the peak denotes the peak hour of ionization as observed from measured hour of ionization as observed from measured data data Figure 6: Relative deviation of modelled and measured VTEC at Figure 7: Diurnal variation of modelled and measured VTEC at Lucknow during 2019. The green and red solid lines denote Hyderabad during 2019. The black, green and red solid lines relative deviation from the measured data while using IRI-plas denote VTEC derived from GPS, IRI-Plas and NeQuick-2 and NeQuick-2 respectively. The vertical black dashed line respectively. The vertical black dashed line denotes the peak denotes the peak hour of ionization as observed from measured hour of ionization as observed from measured data data International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Table 4: Maximum relative deviation and relative deviation at peak hours of ionization between modelled and measured VTEC during 2019 at Hyderabad Month Maximum relative Maximum relative Peak hours of Relative deviation Relative deviation while using deviation while using ionization as during peak hours deviation during IRI-Plas model (%) NeQuick-2 model (%) observed from of ionization while peak hours of measured data using IRI-Plas ionization while January 163.74 at 1300 UT 109.24 at 1300 UT model using NeQuick-2 February 186.54 at 1800 UT 102.80 at 1800 UT 0800 UT model March 248.94 at 1600 UT 237.81 at 1600 UT 0700 UT 89.71 % 45.29 % April 144.41 at 1400 UT 123.04 at 1400 UT 0700 UT 76.46 % 23.71 % May 99.32 at 1100 UT 69.19 at 1200 UT 0800 UT 85.48 % 32.24 % June 104.87 at 1000 UT 55.10 at 1100 UT 0900 UT 81.23 % 26.77 % July 98.77 at 2100 UT 60.89 at 1400 UT 0800 UT 85.72 % 36.58 % August 108.83 at 0900 UT 55.16 at 1500 UT 1000 UT 86.94 % 30.61 % September 127.26 at 1300 UT 89.46 at 1300 UT 0900 UT 82.52 % 30.44 % October 146.16 at 0200 UT 89.82 at 0200 UT 1000 UT 108.83 % 43.29 % November 134.05 at 1400 UT 86.38 at 1400 UT 0800 UT 89.50 % 39.8 % December 129.31 at 0700 UT 73.43 at 1500 UT 0700 UT 86.57 % 34.17 % 0900 UT 83.48 % 34.73 % 109.20 % 65.08 % The negative relative deviation from measured data mostly observed in the post-sunset duration of the is mostly observed in the post-afternoon durations day. Negative relative deviation is also observed of the day, and in some months (January, February between 0000 UT and 0200 UT throughout the year. and December) it is observed between 0000 UT and 0300 UT. Furthermore, GPS data from the IGS From Table 4 it can be observed that, at portal is unavailable for the months of July and Hyderabad, during 2019, the maximum relative November. deviation of modelled VTEC from measured VTEC was observed at different durations of time at From Table 3 it can be observed that, at different months of the year. However, in all months Lucknow, during 2019, the maximum relative of the year except July, August and December, both deviation of modelled VTEC from measured VTEC models show maximum relative deviations from was observed at different durations of time at measured VTEC at approximately the same UT. different months of the year. However, in all months Finally, the maximum relative deviation from of the year except June, August and October, both measured VTEC for the entire year was observed in models show maximum relative deviations from March at 1600 UT for both models. From Table 4 it measured VTEC at the same UT. Finally, the can further be observed that, at Hyderabad, during maximum relative deviation from measured VTEC 2019, the maximum relative deviation of modelled for the entire year was observed in April at 0000 UT VTEC from measured VTEC during peak hours of for both models. From Table 3 it can further be ionization was observed in December while using observed that, at Lucknow, during 2019, the IRI-Plas (109.2%) and NeQuick-2 (65.08%) models. maximum relative deviation of modelled VTEC Negative relative deviation from measured data was from measured VTEC during peak hours of not observed at peak hours of ionization at ionization was observed in March while using IRI- Hyderabad during 2019. Plas (131.81%) and NeQuick-2 (100.95%) models. Negative relative deviation from measured data was 5. Discussions not observed at peak hours of ionization at Lucknow At the mid-latitude region of Lucknow and the low- during 2019. latitude region of Hyderabad, the maximum relative deviation of modelled VTEC from measured VTEC 4.4 Diurnal Variation and Relative Deviation (D) of has a higher magnitude while using IRI-Plas model Modelled and Measured VTEC at Hyderabad for all months of 2015 and 2019. While the During 2019 maximum relative deviation for the entire year at From Figure 7 and Figure 8 it can be observed that, Lucknow was observed in April for both years, the at Hyderabad, during 2019, the relative deviation of same was observed in March at Hyderabad during modelled VTEC from measured VTEC is both 2015 and 2019. During the peak hours of ionization, positive and negative, depending on the duration of when electron content in the ionosphere is at its day. Throughout the year, positive relative deviation highest due to increased solar radiation, the is observed in the earliest hours of the day, and magnitude of relative deviation of modelled VTEC throughout the sunlit hours of the day. Furthermore, from measured VTEC is higher while using IRI-Plas throughout the year, negative relative deviation is model at both regions during 2019. International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Figure 8: Relative deviation of modelled and measured VTEC at Figure 9: Scatter plot of modelled vs measured VTEC at Hyderabad during 2019. The green and red solid lines denote Lucknow during 2015 relative deviation from the measured data while using IRI-plas and NeQuick-2 respectively. The vertical black dashed line denotes the peak hour of ionization as observed from measured data Figure 10: Scatter plot of modelled vs Figure 11: Scatter plot of modelled vs Figure 12: Scatter plot of modelled vs measured VTEC at Hyderabad during 2015 measured VTEC at Lucknow during measured VTEC at Hyderabad during 2019 2019 International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Figure 13: %D at peak hours of ionization for all months of 2015 at Lucknow. Relative deviation of IRI-Plas and NeQuick-2 data from measured VTEC is denoted by diagonally-hatched and horizontally-hatched bar plots respectively Figure 14: %D at peak hours of ionization for all months of 2019 at Lucknow. Relative deviation of IRI-Plas and NeQuick-2 data from measured VTEC is denoted by diagonally-hatched and horizontally-hatched bar plots respectively Figure 15: %D at peak hours of ionization for all months of 2015 at Hyderabad. Relative deviation of IRI-Plas and NeQuick-2 data from measured VTEC is denoted by diagonally-hatched and horizontally-hatched bar plots respectively Figure 16: %D at peak hours of ionization for all months of 2019 at Hyderabad. Relative deviation of IRI-Plas and NeQuick-2 data from measured VTEC is denoted by diagonally-hatched and horizontally-hatched bar plots respectively International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

While the same observation holds true at Lucknow NeQuick-2 model. Furthermore, the magnitude during 2015, the magnitude of relative deviation of relative deviation of modelled data during from measured data has a higher magnitude while peak hours of ionization is higher during solar using the NeQuick-2 model at Hyderabad in the minimum at both regions for both models. months of February, March and April of 2015. Therefore, it can be concluded that, during Furthermore, during peak hours of ionization at both solar minimum, VTEC is highly overestimated regions, negative relative deviation was not by both models during peak hours of observed in any month of 2019. Considerable ionization, and the magnitude of negative relative deviation during peak hours of overestimation at these hours is higher while ionization was only observed in Hyderabad during using IRI-Plas model for both regions. During 2015 while using the NeQuick-2 model. Also, the the solar maximum, VTEC at the mid-latitude magnitude of relative deviation at peak hours of region of Lucknow during peak hours of ionization is higher during 2019 in both regions for ionization showed maximum relative deviation both models, when compared to relative deviation of during the September Equinox and December modelled data at peak hours of ionization during Solstice seasons, while using both models 2015 for the respective regions. (Figure 13). However, during the same year, while VTEC at the low-latitude region of Finally, from Figures 9, 10, 11 and 12 it can be Hyderabad during peak hours of ionization observed that the IRI-Plas data shows higher R2 showed maximum deviation during June value when compared to NeQuick-2 data for both Solstice, September Equinox and December years and at both regions. Furthermore, at both Solstice seasons when using IRI-Plas model, regions, R2 value is higher during 2019. the same was observed while using NeQuick-2 model in the March Equinox season (Figure 6. Conclusions 15). Finally, during solar minimum, At From the above study the following conclusions are Lucknow and Hyderabad, VTEC is highly made: overestimated by both models throughout the year. However, NeQuick-2 model was more • The magnitude of maximum relative deviation accurate in predicting VTEC during peak of modelled VTEC from measured VTEC was hours of ionisation of solar minimum, higher while using IRI-Plas model at Lucknow especially at the low-latitude region of and Hyderabad during all months solar Hyderabad (Figure 15 and Figure 16). maximum and solar minimum. At Lucknow, • The coefficient of determination (R2) values the maximum relative deviation for the entire were high during solar minimum while using year was recorded in April during solar both models at both regions. This implies that minimum and maximum. However, at both models show smaller differences when Hyderabad, the same result was recorded in compared to measured VTEC at both March during solar minimum and maximum. geographic locations during solar minimum. The above-mentioned characteristics of Furthermore, the IRI-Plas model produces a maximum relative deviation at both geographic more accurate representation of VTEC when locations seem to be unaltered by the effects of compared to NeQuick-2 model during solar solar maxima and minima. minimum and maximum. Finally, R2 value at Hyderabad is higher than that at Lucknow • From Figures 13, 14, 15 and 16 it can be during solar minimum and maximum. This summarized that, during the peak hours of indicates that, both models produce electron ionization, magnitude of relative deviation of content at a low-latitude region more modelled VTEC from measured VTEC was accurately, than at a mid-latitude region in higher while using IRI-Plas model for both India. years at both regions, except for three months of 2015 at Hyderabad. The above-mentioned Acknowledgements characteristics of relative deviation during The authors would like to express their gratitude to peak hours of ionization at both geographic the International Multilateral Regional Cooperation locations seem to be unchanged by the effects Division, Department of Science and Technology, of solar maxima and minima. Negative relative Government of India for grant deviation of modelled data during peak hours (IMRC/AISTDF/CRD/2018/000037) towards this of ionization is observed only during solar project on ionospheric studies over Indian maximum at certain months. Considerable subcontinent. The authors extend their gratitude to underestimation of VTEC during peak hours of ionization was only observed in Hyderabad during the solar maximum while using the International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

NASA and IGS for providing the required datasets D’ujanga, F. M., Mubiru, J., Twinamasiko, B. F., to conduct this research. We would like to sincerely Basalirwa, C. and Ssenyonga, T. J., 2012, Total acknowledge the authors responsible for the design Electron Content Variations in Equatorial and implementation of the IRI-Plas model Anomaly Region. Adv. Space Res. Vol. 50, 441- (Gulyaeva et al., 2012, 2013, 2017 and 2018) 449. (Gulyaeva, 2016 and 2019). Similarly, we would like to extend our acknowledgement to the Evans, J. V., 1957, The Electron Content of the Aeronomy and Radio-propagation Laboratory of the Ionosphere. J. Atmos. Terr. Phys., Vol. 11, 259- Abdus Salam International Centre for Theoretical 271. Physics, Trieste, Italy; and the authors (Bruno Nava, Pierdavide Coisson, Johanna Wagner and Yenca Ezquer, R. G., Scida´, L. A., Migoya Orue´, Y., Migoya Orue) responsible for the implementation of Nava, B., Cabrera, M. A. and Brunini, C., 2017, the NeQuick-2 model. NeQuick 2 and IRI Plas VTEC Predictions for Low Latitude and South American Sector. References Advances in Space Research, Vol. 61(7), 1803- 1818. Anderson, A., Anghel, A., Chau, J., Yumoto, K., Battacharyya, A. and Alex, S., 2006, Daytime, Gulyaeva, T. L. and Bilitza, D., 2012, Towards ISO Low Latitude, Vertical ExB Drift Velocities, Standard Earth Ionosphere and Plasmasphere Inferred from Ground-Based Magnetometer Model. In: “New Developments in the Standard Observations in the Peruvian, Philippine and Model”. edited by R.J.Larsen, 1-39, NOVA, Indian Longitude Sectors Under Quiet and Hauppauge, New York, [Available at https://w- Disturbed Conditions. ILWS Workshop, Goa. ww.novapublishers.com/catalog/product_info.ph 19-24. p?products_id=35812] Appleton, E. V., 1946, Two Anomalies in the Gulyaeva, T. L., Arikan, F., Hernandez-Pajares, M., Ionosphere. Nature, Vol. 157, 691-693. Stanislawska, I., 2013, GIM-TEC Adaptive Ionospheric Weather Assessment and Forecast Bauer, S. J. and Daniels, F. B., 1959, Measurements System. J. Atmosph. Solar-Terr. Phys., Vol. 102, of Ionospheric Electron Content by the Lunar 329-340, doi:10.1016/j.jastp.2013.06.011. Reflection Technique. J. Geophys. Res., 64, 1371- 1376. Gulyaeva, T., 2016, Modification of the Solar Activity Indices in the International Reference Bhattacharya, S., Purohit, P. K. and Gwal, A. K., Ionosphere IRI and IRI-Plas Models Due to 2009, Ionospheric Time Delay Variations in Recent Revision of Sunspot Number Time Equatorial Anomaly Region During Low Solar Series. Solnechno-Zemnaóa Fizika. Vol. 2(3), Activity Using GPS. Indian J. Radio Space 59-68. DOI:10.12737/20872. [Available at: htt- Phys., Vol. 38, 266-274. p://ru.iszf.irk.ru/Journal_Solar-Terrestrial- _Physics._Vol._2%2C_Iss._3%2C_2016 (in En- Bilitza, D., 1990, International Reference glish)]. Ionosphere 1990, NSSDC/ WDC-A-R&S 90-22. National Space Science Data Center, Greenbelt. Gulyaeva, T. L., Arikan, F., Poustovalova L., Sezen U., 2017, TEC Proxy Index of Solar Activity for Bilitza, D., Radicella, S., Reinisch, B., Adeniyi, J., the International Reference Ionosphere IRI and Mosert, M., Zhang, S. and Obrou, O., 2000, New its Extension to Plasmasphere IRI-PLAS Model. B0 and B1 Models for IRI. Adv Space Res, Vol. Int. J. Sci. Eng. Applied Sci., Vol. 3(5),144-150, 25(1), 89-95. http://ijseas.com/index.php/issue-archive-2/vol- ume3/issue-5/. Bilitza, D. and Reinisch, B. W., 2008, International Reference Ionosphere 2007: Improvements and Gulyaeva, T. L., Arikan, F., Sezen, U. and New Parameters. Adv Space Res., Vol. 42(4), Poustovalova, L. V., 2018, Eight Proxy Indices 599-609. doi:10.1016/j.asr.2007.07.048. of Solar Activity for the International Reference Ionosphere and Plasmasphere Model. J. Atmos. Chauhan, V. and Singh, O. P., 2010, A Solar-Terr. Phys., Vol. 172, 122-128, Morphological Study of GPS-TEC Data at Agra https://doi.org/10.1016/j.jastp.2018.03.025. and their Comparison with the IRI Model. Adv. Space Res., Vol. 46, 280-290. Gulyaeva, T. L., 2019, Predicting Indices of the Ionosphere Response to Solar Activity for the David, H. H., 2015, The Solar Cycle. Living Rev. Ascending Phase of the 25th Solar Cycle. Adv. Solar Phys., Vol. 7(1), 12. Space Res., Vol. 63,(Is.5), 1588-1595, https://doi.org/10.1016/j.asr.2018.11.002. Dieter, B., Lee-Anne, M., Bodo, R. and Tim, F., 2011, The International Reference Ionosphere Gulyaeva, T. L., 2020, IRI-Plas (version IRI-Plas- Today and in the Future. J. Geod, Vol. 85, 909- 2017) [Source code]. ftp://ftp.izmiran.ru/pub- 920. /izmiran/SPIM/. International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Kenpankho P., Watthanasangmechai, K., Supnithi, Sandro, R., 2009, The NeQuick Model Genesis, P., Tsugawa, T. and Maruyama, T., 2011, Uses and Evolution. Annals of Geophysics, Vol. Comparison of GPS TEC Measurements with 52(3/4), 417-422. IRI TEC Prediction at the Equatorial Latitude Station, Chumphon, Thailand. Earth Planets Tariku, Y. A., 2015, TEC Prediction Performance of Space, Vol. 63, 365-370. IRI-2012 Model During a Very Low and a High Solar Activity Phase Over Equatorial Regions, Kouba J., 2009, A Guide to Using International Uganda. J. Geophys. Res. Space Physics, Vol. GNSS Service (IGS) Products. Natural 120, 5973-5982. Resources Canada. 1-34. Unnikrishnan, K., Balachandran, R. and Venugopal, Mendillo, M., 2005, Storms in the Ionosphere: C., 2002, A Comparative Study of Nighttime Patterns and Processes for Total Electron Enhancements of TEC at Low Latitude Station Content. Rev. Geophys., Vol. 44, RG4001. on Storm and Quiet Nights Including the Local Time, Seasonal and Solar Activity Dependence. Okoh, D., Mckinnell, L., Cilliers, P. and Okere, B., Ann. Geophys. Vol. 20, 1843-1850. 2015, IRI-vTEC Versus GPS-vTEC for Nigerian SCINDA GPS Stations. Advances in Space Zhang, M. L., Radicella, S. M., Leitinger, R., Nava, Research, Vol. 55(8), 1941-1947. B., Coisson, P., Wagner, J. and Orue, Y. M., 2010, Aeronomy and Radiopropagation Sandro, R. and Man-Lian, Z., 1995, The Improved Laboratory of the Abdus Salam International DGR Analytical Model of Electron Density Centre for Theoretical Physics, Trieste, Italy Height Profile and Total Electron Content in the (2010) NeQuick-2 (version 2.0.2) [Source code]. Ionosphere. Annals of geophysics, Vol. 38(1), https://t-ict4d.ictp.it/nequick2/source-code. 35-41. International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Geoinformation Modeling of Knowledge Spillovers as an Innovative Development and Agricultural Efficiency Factor Nosonov, A.,1 Letkina, N.2 and Nosonova, V.3 1National Research Mordovia State University, Department of Physical and Socio-Economic Geography, Russia, E-mail: [email protected] 2Department of English for Professional Communication, Faculty of economics, Department of Public and Municipal Administration, Russia, E-mail: [email protected] 3Saransk, Russia, E-mail: [email protected] Abstract The article is devoted to the study of spatial regularity of knowledge spillovers in agriculture as an innovative development factor based on the use of geoinformation modeling methods. The spatial localization of patents development in the agricultural sector and the main regions of their citations have been analyzed which characterizes the secondary effects of knowledge dissemination. It was found that references to agricultural patents at the initial stage in most cases coincide with the areas of new knowledge generation. During the subsequent period this narrow localization of citations disappears and their distribution area expands significantly. The use of GIS technologies and geoinformation modeling allowed us to visualize the process of knowledge spillovers in the agricultural sector and to reveal the spatio-temporal patterns of agricultural innovative knowledge dissemination in various regions of Russia. The influence of knowledge spillovers in agriculture on the efficiency of agricultural production has been analyzed for the first time. We have discovered the dependence of patent activity in agriculture on the general level of regions innovative development. A typology of Russian regions on innovative functions based on the ratio of registered and used patents in agriculture was carried out. In accordance with this three types of regions have been distinguished – creative, acceptor-creative and acceptor. Within the identified types of regions, differences in the efficiency of agriculture were revealed that testify their close dependence both on the ratio of creative and acceptor functions and on the general level of regions innovative development. 1. Introduction agriculture is aimed at a more efficient use of the The current stage of society development is territory natural agricultural potential, socio- characterized by the formation of an economy based economic factors, institutional conditions which can on knowledge, skills and technology as a result the be achieved only on the basis of innovative high level of economic development of the world development. Reserves of extensive agricultural leading countries is largely determined by the growth were depleted in the late 1950s. effective integration of science, education and Subsequently the entire increase in agricultural business. Currently, successful development of any production was achieved only through country is determined by the level and the degree of intensification which is based on the use of new use of the results of scientific activities implemented equipment and technologies, genetics development, on the market in new equipment and technologies breeding and genetic engineering, the expansion of i.e. the commercialization of the innovation results melioration and farmland chemicalization, (Baburin and Zemtsov, 2017 and Zamyatina et al., organizational and managerial innovations, and the 2020). The production and new knowledge improvement of human potential in the agricultural spillovers, the results of scientific research from a sector. Nowadays the threat of ensuring food social resource turn into an element of market security of the country has increased as a result of relations, into the mechanism of the country's changes in the main areas of foreign economic trade competitive struggle for leadership in the field of policy in the field of agro-food products under the innovative development. In this sense, science is sanctions of a number of countries supplying realized as part of the economy innovative agricultural products and agricultural machinery component (Makar and Nosonov, 2017). products. This generates a need for adjustments to In Russia at the present stage of development International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

the national agricultural policies in the direction of of knowledge is a process of interaction between expanding their own food production on an individuals in the course of which there is a innovative basis. State and private financing can conversion of implicit knowledge into explicit and provide food import substitution aimed at increasing vice versa (Figure 1) (Nonaka and Takeuchi, 1995). and qualitative improvement of means of production Implicit knowledge is the most significant for and equipment, at developing the production and innovation. This is informal knowledge which social infrastructure and human resource represents individual knowledge, skills and development. The implementation of the Food professional experience and can only be transmitted Import Substitution Federal Program over the past verbally through personal contacts in a very limited four years has significantly improved country’s food epistemological community, for example, the security which was critical in a number of areas. At transfer of knowledge from a teacher to a student. the present moment, Russia has fully ensured its Explicit (codified) knowledge is a formalized and food security for most types of food which is no less systematized knowledge that is easily transmitted by important than military-strategic or economic using paper and electronic media in the form of stability. From 2013 to 2018 the ratio of imports and books, scientific articles, formulas, etc. exports of food products and agricultural raw materials has significantly changed. Over this Innovations arise in the process of transforming period, food exports increased from $16 billion in different types of knowledge. The following stages 2013 to $25 billion in 2018, the import rate of these of knowledge transformation are distinguished as a products decreased from $43 billion in 2013 to $30 result of their transfer (Nonaka and Takeuchi, 1995 billion in 2018 (Regions of Russia, 2018). and Fischer, 2001): The study of agricultural innovation 1. from implicit to explicit knowledge development trends based on the research of (externalization). This is the process of innovation diffusion as well as including the creating new conceptual models by codifying analysis of intra-industry knowledge spillovers is an implicit knowledge. understudied problem, the significance of which 2. from explicit knowledge to implicit increases significantly in the conditions of mobile (internalization). This is a first-hand training and dynamic exchange of scientific, technical and based on formal instructions. technological information between regions and 3. from implicit to implicit knowledge countries. (socialization). 4. from explicit to explicit knowledge 2. Methodology (combinations). This is a systematic approach 2.1 Literature Review to the creation of new knowledge on the basis At present, there is a tendency to increase the of heterogeneous types of codified knowledge, number of scientific studies devoted to identifying which is used in training new employees, as and studying the factors of countries and regions well as during monitoring visits and innovative development which is considered to be instructions. the main source of their economic growth. Knowledge and its dissemination in time and space, There are several approaches to the analysis of i.e., knowledge spillovers take an important place knowledge spillovers (Feldman, 1999): among these factors. The knowledge spillovers is understood as the process of transferring knowledge Modified Production Function of Knowledge: created in one company or research center to taking into account the territorial scope and another company free of charge or with little characteristics of the created innovations (Griliches, compensation, much less than the cost of knowledge 1992 and Jaffe, 1989). itself (Zamyatina et al., 2020 and Romer, 2015). According to Japanese scientists the transformation Isi = IRDsi * URsi * εsi Equation 1 Figure 1: Four main transforming knowledge processes (Nonaka and Takeuchi, 1995) International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

where I - innovative return; IRD – private analysis of discrete processes and phenomena is expenditures of corporations on research and especially high when the set of points with the technological development (RTD); UR – University initial spatial data is dense enough to reveal the RTD expenses. degree of local change in the displayed surface. Considering that the IDW interpolation is used to Analysis of Patent Activity and Patents Citation: study the following phenomena and processes: Some economists believe that it is impossible to trace the «paper traces» of knowledge spillovers - geo-ecological problems associated with the (Krugman, 1991). However, Jaffe et al., (2000) have pollution of point objects (Matějíček et al., developed an approach that involves the analysis of 2006, Anjusha et al., 2020 and Meng, 2020). patents with its references to other patents (i.e. patent citation), which is the basis for the analysis of - designing of digital elevation models both territorial and temporal aspects of knowledge (Pellitero et al., 2016, Soleimamani and spillovers (Jaffe et al., 1993 and 2000). The Modallaldoust, 2008). relationship between the patent’s place of creation and the territories of its citation can be used to - research of negative social phenomena identify spatial knowledge spillovers. A spatial (Ansari and Kale, 2014 and Achu and Rose, pattern has been revealed that the number of 2016). references to patents (as an expression of explicit (codified) knowledge) significantly decreases with - spatial analysis of emerging infectious increasing distance, which indicates the decisive diseases (Blanco et al., 2019, Samadzadeh et role of spillovers of implicit knowledge. al., 2019 and Belief, 2018). The Intellectual Capital of the Region (Ideas - analysis of climatic indicators (Earls and Embodied In People): Many studies show that the Dixon, 2007 and Wang et al., 2010). main mechanism for disseminating knowledge at the local level is the transfer of scientists, engineers, and Despite the widespread use of spatial interpolation innovation managers both between companies and methods in the study of natural, economic, social between firms and universities, research institutes and environmental problems, their insufficient use and vice versa. Currently, the dissemination of for the analysis of innovative processes and knowledge is mainly due to personal contacts and phenomena should be eliminated. However, the the mobility of innovators. The main carriers of possibilities of spatial interpolation methods, in knowledge are people who are as a rule familiar, particular IDW interpolation, can effectively solve trust each other and share acquired knowledge. such problems of innovative development of There is also an indirect transfer of knowledge countries and regions as the study of innovations through scientific articles and as a result of scientists diffusion and the spillover of knowledge, identify views exchange the scientific conferences (Breschi factors stimulating innovation, conduct a typology and Lissoni, 2001). of regions on innovative functions, etc. Similar studies are necessary for organizations and decision Innovations implemented in goods and services and makers to justify managerial impacts on socio- Related to International Trade (Ideas Embodied in economic systems in order to increase the efficiency Goods): This is manifested in international RTD of their functioning. spillovers and high technologies transfer. The dissemination of knowledge in this case is carried The results of spatial interpolation of knowledge out through the patents sale, obtaining licenses for spillovers are the basis for identifying the types of inventions and useful models, and the supply of regions according to the ratio of creativity and high-tech equipment and technologies. acceptance. There are many approaches to highlight the level of regions creativity. According to In this study we have used a methodology to Schumpeter (2003) the indicator of community analyze patent activity and citation of patents. We creativity is the share of innovators (inventors, have also used the IDW interpolation method for the entrepreneurs) in it (Schumpeter, 2003). A classical first time to visualize the process of knowledge approach is that proposed by Richard Florida which spillover. That method is widely used in geo- is based on the identification of employment areas. information modeling of geographical objects, R. Florida unites people engaged in creative phenomena and processes due to its high efficiency. professions under the general term \"creative class\" The effectiveness of the method in the spatial or the class of creative professionals. These include scientists, engineers, architects, programmers, educators, and designers, representatives of the arts, entertainment, sports and mass media (Florida, 2002, 2005 and 2006). The creative class makes the maximum contribution to the GDP growth of developed countries (Florida, 2002). The greatest International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

concentration of creative professionals is with a higher average density of innovations); characteristic mainly in developed Western acceptor-creative (with a higher average generation countries, as well as in the capitals and major cities of innovations, but with a high share of inventions of developing countries and countries with used, exceeding 100%); strong acceptors (with a economies in transition. According to R. Florida relatively low generation of innovations, but with a (Florida, 2006) the creativity of the regions is very high share of inventions used, exceeding determined by the share of the creative class 100%); weak acceptors (with low generation of concentration, inventive activity and the diversity of innovations and with the share of used inventions the local community. He developed a creativity above the average level, but less than 100%); index based on the concept of three \"T\": talent, innovation periphery (with extremely low rates) technology and tolerance. An approach based on a (Baburin, 2010). modification of the methodology of Florida (2002) was used in the studies of Pilyasov and Kolesnikova 2.2 Methodology and Research Methods (2008). The approach is based on the assumption GIS technologies are not sufficiently used in the that the more talented, technological, and tolerant a study of spatial knowledge spillovers which allow community is, the more creative it is. The indicators you to visually display the distribution of of creativity index are similar to those proposed by knowledge in the form of patent citations across the R. Florida (Florida, 2005); for those absent in territory. Currently, two main approaches are Russian statistics, the equivalents are used: the share applied to the integration of geographic information of employees with higher education, the number of technologies and models of socio-economic researchers per 1 million inhabitants, the share of (including innovative) processes in order to create a R&D expenses in GRP, the number of patents spatially distributed modeling system (Kapralov et granted per million inhabitants and others. Common al., 2008, Lurie, 2008 and Teslenok et al., 2014). to these approaches is the determination of regions The first one uses the GIS package software as an creativity by the concentration of the creative class additional block of the computer model of the in them, including innovators and inventors. process providing the formation of input data arrays and the presentation of simulation results performed Baburin (2010) and Baburin and Zemtsov by traditional methods (Manju et al., 2019). The (2017) proposed two approaches for determining the second approach involves full integration of GIS level of creativity of regions. The first approach and a process profile model implemented by the takes into account the dependence of the number of capabilities of the GIS software package. connections and innovations on the size of the urban population and the density of cities. The higher the The source of information for the study of density of cities and the urban population, the higher knowledge spillovers in agriculture was the the innovative and creative potential of the region. bibliometric data on the number of patents and their The second approach is based on different creative citations for agriculture, which are contained in the and acceptor abilities of the community, depending databases of the Scientific Electronic Library on different ages in the demographic structure. eLIBRARY.RU for 2010-2019. (12,670 patents and Younger people (0-25 years old) most actively 51248 citations) (Scientific electronic library, 2019). accept information. In the future, at the age of 25- Design work on the possible options and the 50, they become the main reinventors and development of their database structure based on innovators supporting and disseminating social and GIS ArcView GIS by ESRI, Inc. has been carried technological inventions. In their 50s people out and it became the basis for geoinformational become a conservative element of society, giving it modeling and mapping. Based on the nature of the stability. Based on a generalization of these two initial statistical data the number, general list, names approaches, all regions of the country can be and attribute tables field parameters of the designed attributed to the regions that create new GIS have been determined. technologies (creative «donors») and the regions that consume technologies («acceptors»). To assess After creating the corresponding new project in the first factor, an indicator of the number of patents the GIS ArcView the themes (layers) outlined at the per 100 thousand urban residents is used, as a design stage have been formed, the project and its second indicator - ratio of the share of embedded layers have been configured, the fields with patents to created ones. In accordance with this, previously defined parameters have been formed in Baburin (2010) identified the following types of the attribute tables of the respective topics. The regions in terms of creativity: creative (using source data for inclusion in the GIS database are inventions much less than they are created); sub- presented in Excel format (*.xls). Preliminary, the creative (using inventions less than they produce; files (books) with data in the context of the territorial entities of the Russian Federation had International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

been created in Microsoft Excel for each of the the potential need for further frequent use, the analyzed indicators. Then, the active sheets of all legends obtained can be saved in the corresponding books were saved in *.dbf (DBF 4 (dBASE IV)) files (*.avl). To study the main directions of spatial format in order to provide the possibility of further spillovers of agricultural knowledge, the work with them in ArcView GIS since the shapefile interpolation method between the centers for attributes are stored in this format. creating patents and the regions of their citation was applied. For these purposes the Inverse Distance The classification of Russia territorial entities by Weighted Interpolation (IDWI) method was used the number of patents granted and their citations and based on the weighting of points in such a way that visual presentation of the results in the form of the influence of a known value decays with cartograms were performed using the GIS legend increasing distance to an unknown point, the editor ArcView GIS and the type of legend «color parameters of which must be determined. Weighing scale», when changing the values of the attribute is assigned by the source centers with a different data of the theme objects is represented by the range number of patents using a weighting coefficient that of the spectrum of one color scale with the initial estimates how the influence of the parameter will and final colors. decrease with increasing distance to it. The higher the value of the weighting coefficient the less will By default, the capabilities of the base GIS be the effect exerted by the point. As the coefficient ArcView GIS allow automatic classification of increases the value of the unknown point will mapped objects using the «color scale» legend type approach the value of the nearest point with a in four ways (natural boundaries, equal intervals, known value. fractiles, standard deviations) according to a numerical attribute – a field with an analyzed To study the main directions of spatial flows of indicator. agricultural knowledge, we have applied the interpolation method between the centers for The method (type of classification) of natural creating patents and the regions of their citation. For boundaries (intervals) is used by default in ArcView these purposes we have usedthe inverse-distance GIS. It defines the boundary points between classes weighting (IDW) method. This method is based on (the so-called breakpoints) using the statistical the weighting of points in such a way that the Jenk’s optimization formula. The method is based influence of a known value decays with increasing on minimizing the sum of values deviations within distance to an unknown point, the parameters of each class which allows to group data close in which must be determined. Weighing is assigned by values. Classification by the method of equal the source centers with a different number of patents intervals (equidistant) divides the general range of using the weighting coefficient, which estimates attribute values into equal-sized subranges on which how the influence of the parameter will decrease theme objects are then distributed. The method of with increasing distance to it. The higher the value quantiles (uniform, equal) allows you to include the of the weighting coefficient, the less will be the same number of objects in each class and is most effect exerted by the point. As the coefficient suitable for linearly increasing data that does not increases, the value of the unknown point will have a disproportionate number of objects with the approach the value of the nearest point with a same values. In the above mentioned classification known value. IDW interpolation is used when the methods the possible sequence of colors of the mapping phenomenon (process) is characterized by legend scale from initial to final when constructing localization at individual points and their it is determined by the order of the colors in the quantitative characteristics naturally decrease when continuous spectrum (red, orange, yellow, green, moving from the center to the periphery. The blue, blue, violet). In this study the method of equal phenomenon being mapped (knowledge spillovers) intervals (equal to intermediate) turned out to be fully complies with these requirements. optimal using which all values were divided into three equal-sized groups in accordance with the ESRI ArcVIEW defaults to 12 interpolation value of the analyzed indicator: high, medium and methods: Inverse Distance to a Power; Kriging low. Minimum Curvature; Modified Shepard’s Method; Natural Neighbor; Nearest Neighbor; Polynomial In the future for better visual perception, greater Regression; Radial Basis Function; Triangulation visibility and ease of use when analyzing the results with Linear Interpolation; Moving Average; Data of classification also using the legend editor, the metrics, Local Polynomial. During the research all colors of the intervals of values of each class can be 12 interpolation methods were tested. The selected and a color ruler (color change scale) with a specificity of the spatial data used and the analysis transition from dark colors of one gamma of the results obtained made it possible to select the corresponding to high the level of the indicator to light, indicating a low level of the parameter. Given International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

most optimal method, Inverse Distance to a Power, technical, technological, organizational and using the expert evaluation method. managerial knowledge through diffusion of innovations, etc.); At the final stage of the study a typology of the – widespread of organizational and managerial regions of Russia on their creative and acceptor innovations in the agro-industrial complex – functions was carried out, based on the ratio of foundation of agricultural holdings and other issued and used patents for inventions and useful vertically oriented management structures. This type models in the agricultural sector which reflects of innovation due to the high territorial innovative functions of country’s regions. When concentration of labor, capital and cooperation is the performing this typology we have used the focus of a wide range of innovations - technical, methodology proposed by Baburin (2010) which technological, marketing, social, etc. Promising was modified and adapted as applied to agriculture. innovations that are at the research and development The main classification feature is the number of stage are being tested in agricultural holdings: patents and their citations, as well as the ratio of agricultural robots; closed ecological systems; invented and used patents. The following indicators genetically modified food; vertical farms. A special were used as additional indicators: the general level place in modern agrarian innovations is occupied by of region’s innovative development, the ratio of precision farming technology, which takes into developed and used production technologies, account the peculiarities of the topsoil, microclimate internal costs of scientific research and development and topography within the boundaries of individual in agricultural sciences, the graduation of bachelors fields. To determine and evaluate these local and masters of agricultural specialties, number of differences, high-tech methods are used – personnel engaged in agricultural research and GLONASS global positioning, geo-information development (including those with a PhD degree), systems and technologies for remote sensing of the etc. In accordance with this three groups of regions Earth, information technologies of agro- are distinguished: creative, in which much less management, methods of technological inventions are used than are created; acceptor, standardization, etc. which are characterized by low generation of In agriculture, increased costs on IRD and innovations and a high proportion of inventions and improving knowledge dissemination mechanisms utility models used; acceptor-creative, occupying an are gaining great importance. An analysis of patent intermediate position. A separate group consists of activity shows that four main centers have been regions in which all indicators of innovative formed in Russia with the largest number of created development are extremely low - the innovation patents for inventions and utility models: Moscow periphery. Within the selected types a comparative and the Moscow Region, St. Petersburg, the North analysis of the main indicators of agricultural Caucasus and Chernozemye and the Middle Volga efficiency was carried out to identify the impact on Region (Figure 2). This circumstance is due to the these parameters of patent activity and knowledge fact that a large number of both research institutes spillovers in the agricultural production system. and agricultural centers, and agricultural higher education institutions is concentrated in these 3. Results and Discussion regions. Also in these constituent entities of the Modern agriculture is an extremely complex and Federation work the most qualified specialists in the poorly structured system therefore when studying agricultural sector and there are opportunities the knowledge spillovers in this field it is necessary (machinery and equipment) for doing research work to apply scientific approaches and methods that take in the high-tech spheres of the agricultural sector – into account the following characteristics of that genetics, breeding, agrobiotechnology, genetic material production sphere: engineering, and the production of nano- and – a significant impact on agriculture of natural composite materials. The number of agricultural patents citations is conditions and territory resources and the geographically limited (Figure 3). Their largest presence of natural cyclicality; number falls on the following regions of Russia: – a longer period of using fixed assets compared to Moscow and Moscow Region (8784 citations), industry; Ulyanovsk Region (5599), Ryazan Region (4796), – the existence of long-term agrarian crises and a St. Petersburg and Leningrad Region (4344), significant period of rehabilitation; Krasnodar Territory (2917) and Stavropol Territory – significant dependence on organizational and (2444). An analysis of the actual knowledge production innovations (systems evolution of spillovers in agriculture (Figure 3) indicates their land use, degree of mechanization of production very limited localization within the main centers for processes, reclaimed land area, mineral and organic fertilizer application, spillovers of International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

generating agricultural innovations. Figure 2: The number of patents in agriculture (a) and the number of patents citations in agriculture (b) registered in the Scientific Electronic Library in 2010-2019, units. Compiled by (Scientific electronic library, 2019) Figure 3: The interpolation by the inverse distance wights method of patents number in agriculture and their citations («knowledge spillovers»). Compiled by (Scientific electronic library, 2019) Figure 4: Types of agricultural regions in terms of creativity: 1 -creative; 2 – acceptor-creative; 3 – acceptor. Compiled by the authors International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

There are only two large areas of knowledge agricultural innovations generated in creative spillovers – within Moscow and the Moscow region regions. Regions of an acceptor-type and innovative ,and in the North Caucasus within the Stavropol and periphery occupy territories with low natural Krasnodar territories. Less significant areas of resource potential, insignificant labor supply and knowledge flow in agriculture were also formed dominance in the sectoral structure of industrial around St. Petersburg, Kazan, Penza, Ulyanovsk, production and services. Voronezh, Omsk and Novosibirsk. One of the study objectives was to analyze the On the territory of the Russian Federation, three differences in the efficiency of agricultural types of regions are distinguished according to the production in different types of regions depending ratio of creative and acceptor innovative functions on the general level of innovative development and of agriculture: creative, acceptor-creative and patent activity (Table 1). The following indicators acceptor (Figure 4). The boundaries between the were used as criteria for the economic efficiency of distinguished types generally correspond to agriculture: the balanced financial result of gradations in the number of registered patents in agriculture (profit-loss) (USD), agricultural agriculture, and are also well manifested on the map profitability (%) and the value of gross agricultural of knowledge spillovers (Figure 3). In nine regions output per 1000 ha of agricultural land (USD/ha). A of the creative type, 55% of patents were created in regular decrease in the efficiency of agriculture the field of agriculture from the all-Russian level during the transition from creative to acceptor and the proportion of their citations is 57%. This is regions was revealed. more than in the remaining 74 regions of the country. The citation rates for patents are also high: 4. Conclusion from 2.9 citations per patent in acceptor regions to Based on the results of the study we can draw the 4.2 in creative ones. Creative-type regions were following conclusions. formed under the influence of the following factors: firstly, high scientific and technical capacity and the 1. The knowledge spillovers (especially explicit) availability of appropriate infrastructure for in agriculture has not yet become a decisive factor agricultural research (Moscow and Moscow Region, in the industry innovative development. In the future St. Petersburg); secondly, the high natural this will be facilitated by the development of agricultural potential of the territory and a good information technologies (the availability of supply of labor resources (North Caucasus and the computers and mobile devices, access to the Middle Volga region). In nine creative regions more Internet, the expansion of a range of electronic than 80% of all agricultural research institutes and services provided, including social ones). Thus, about 70% of higher education institutions specialists of the agricultural sector have increased conducting agricultural research the country are opportunities for improving their level of concentrated. Acceptor-creative regions are located qualification which is the basis for the quality within the main agriculturally developed territory of growth of human potential as an important factor in European Russia and are the main consumers of innovative development. Table 1: Key indicators by region type Main characteristics Types of regions creative creative acceptor acceptor Regional Innovation Index (by HSE rating) 0,438 0,355 0,306 The number of patents for inventions and utility models, units, total 6922 4426 1322 on average per region 769 192 26 The number of citations of patents for inventions and utility models, units total 29320 18065 3863 on average per region 3257 785 77 Balanced financial result of agriculture, million USD* 2645 1598 496 Profitability of agriculture,%* 10.3 9.4 1.6 Agricultural products per 1000 ha of agricultural land, 43,4 22,0 17,3 thousand USD / ha* Crop yield, cwt / ha* 41,8 23,5 19,8 Number of regions in type 9 23 50 Compiled by (Regions of Russia, 2018, Scientific electronic library, 2019, Abdrakhmanova et al., 2017) * average type International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

2. The highest patent activity in the agrarian E., Timofeev, A., Tochilina, E., Fridlyanova, S. sector is observed in the regions with developed and and Fursov, K., 2017, Russian Regional efficient agriculture (the North Caucasus and the Innovation Scoreboard/L. Gokhberg (ed.); Volga region) and in two «capitals» with adjacent National Research University Higher School of areas – Moscow and St. Petersburg where a highly Economics. Issue 5, Moscow: HSE, 260. efficient agribusiness has been also formed. Achu A. L., Rose, R. S., 2016, GIS Analysis of Crime Incidence and Spatial Variation in 3. Unlike the hierarchical mechanism of Thiruvananthapuram City, International Journal dissemination of innovation diffusion in most of Remote Sensing Applications, Vol. 6, 1-7. sectors of the economy, in agriculture this process Anjusha, K. V., James, A. M., Thankachan, F. A., has a more network nature. This is due to the lack of Benny, J. and Hezakiel, V. B., 2020, pronounced centers for the generation of agricultural Assessment of Water Pollution Using GIS: A innovations in the country and a significant Case Study in Periyar River at Eloor Region. territorial dispersal of their centers and regions. Green Buildings and Sustainable Engineering. 413-420. 4. The creation of various agricultural Ansari S. M. and Kale K. V., 2014, Mapping and innovations is significantly affected by Analysis of Crime in Aurangabad City Using environmental conditions which determine the GIS. IOSR Journal of Computer Engineering. different nature of innovations (crop or livestock). Vol. 16(4), 67-76, DOI:10.9790/0661-16476776. The generation of high-tech innovations (in the field Baburin, V. L. and Zemtsov, S. P., 2017, Innovative of genetics, breeding, genetic engineering, etc.) Potential 0f Russian Regions. Moscow: KDU, related to the fourth technological mode is 358. concentrated in large innovation centers at the Baburin, V. L., 2010, Innovation Cycles in the national level. Russian Economy. Moscow: KRASAND, 216. Belief, E., 2018, GIS Based Spatial Modeling to 5. Large territorial entities in the agro-industrial Mapping and Estimation Relative Risk of complex (agricultural holdings) provide a real Different Diseases Using Inverse Distance technological breakthrough by continuous updating Weighting (IDW) Interpolation Algorithm and of the material-and-technical part of agriculture and Evidential Belief Function (EBF)(Case study: introducing innovations: high-tech information Minor Part of Kirkuk City, Iraq). International systems for agricultural management, advanced soil Journal of Engineering & Technology, Vol. cultivation technologies, efficient use of mineral 7(4.37), 185-191. fertilizers and pesticides, information technologies Blanco, I., Diego, I., Bueno, P., Fernández, E., for production management and GIS technology Casas-Maldonado, F., Esquinas, C. and applications. Miravitlles, M., 2019, Geographic Distribution of Chronic Obstructive Pulmonary Disease 6. The obtained results show a significant impact Prevalence in Africa, Asia and Australasia. The on the agriculture innovative development of patent International Journal of Tuberculosis and Lung activity and the general level of regions innovative Disease, Vol. 23(10), 1100-1106. development which are currently the main factor in Breschi, S. and Lissoni, F., 2001, Knowledge increasing the efficiency of agricultural production Spillovers and Local Innovation Systems: A and increasing the production of food resources. Critical Survey, Industrial and Corporate Change. Vol. 10(4), 975-1005. Acknowledgment Earls, J. and Dixon, B., 2007, Spatial Interpolation The study was carried out with support from the of Rainfall Data Using ArcGIS: A Comparative Russian Foundation for Basic Research within the Study. In Proceedings of the 27th Annual ESRI research project № 19-05-00066. International User Conference. Vol. 31. 1-9. Feldman, M. P., 1999, The New Economics of References Innovation, Spillovers and Agglomeration: Areview of Empirical Studies. Economics of Abdrakhmanova, G., Bakhtin, P., Gokhberg, L., Innovation and New Technology. Vol. 8(1–2), 5- Ditkovskiy, K., Islankina, E., Kindras, A., 25. Kovaleva, G., Kovaleva, N., Vera, K., Fischer M. M., 2001, Innovation, Knowledge Kuznetsova, I., Kuzmin, G., Kuzminov, I., Creation and Systems of Innovation. The Kutsenko, E., Martynov, D., Martynova, S., Annals of Regional Science. Vol. 35(2), 199- Nechaeva, E., Ratay, T., Sagieva, G., Streltsova, 216. International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Florida, R., 2002, The Rise of the Creative Class: Taihu Sediments. Taihu Sediments (Doctoral And How It’s Transforming Work, Leisure, Dissertation, NUI Galway. 133. Community and Everyday Life. Canadian Nonaka, I. and Takeuchi, H., 1995, The Knowledge- Public Policy, Vol. 29(3) 1-10. Creating Company. How Japanese Companies Create the Dynamics of Innovation. New York, Florida, R., 2005, Cities and the Creative Class. City Oxford. Oxford University Press, 46-49. & Community, Vol. 2(1), 3-19. Pellitero, R., Rea, B. R., Spagnolo, M., Bakke, J., Ivy-Ochs, S., Frew, C. R. and Renssen, H., 2016, Florida, R., 2006, The Flight of the Creative Class: GlaRe, a GIS Tool to Reconstruct the 3D The New Global Competition for Talent. Liberal Surface of Palaeoglaciers. Computers & Education, Vol. 92(3), 22-29. Geosciences, Vol. 94, 77-85. Pilyasov, A. N. and Kolesnikova, O. V., 2008, Griliches, Z., 1992, The Search for R&D Spillovers. Evaluation of Creativity of the Russian Regional Scandinavian Journal of Economics. Vol. 94, Communities, Economic Issues, Vol. 9, 50-69. 29-47. Regions of Russia, 2018, Socio-Economic Indexes: Stat. Digest, Moscow, 1162. Jaffe, A. B., Trajtenberg, M. and Fogarty M. S., Romer, P. M., 2015, Mathiness in the Theory of 2000, Knowledge Spillovers and Patent Economic Growth, American Economic Review. Citations: Evidence from a Survey of Inventors. Vol. 105(5), 89-93. American Economic Review. Vol. 90, 215-218. Samadzade, R., Habibzadeh, S., Mohammadshahi, DOI: 10.1257/aer.90.2.215. J., Shokrollahi, M. and Moradi-Asl, E., 2019, Spatial Epidemiology of tuberculosis in Ardabil Jaffe, A., 1989, The Real Effects of Academic Province: Based on Geographical Information Research. American Economic Review. Vol. System, Pakistan Journal of Medical and Health 82(1), 363-67. Sciences, Vol. 13(3), 965-969. Schumpeter, J., 2003, Capitalism, Socialism and Jaffe, A., Trajtenberg, M. and Henderson, R., 1993, Democracy. London & New York: Taylor & Geographic Localization of Knowledge Francis e-Library, 460. Spillovers as Evidenced by Patent Citations. Scientific electronic library eLIBRARY.RU. 2019. Journal of Economics. Vol. 63, 577-98. Available: https://elibrary.ru/ Soleimamani, K. and Modallaldoust, S., 2008, Kapralov, E. G., Koshkarev, A.V. and Tikunov, V. Production of Optimized DEM Using IDW S., 2008, Geoinformatics / Ed. by V. S. Tikunov. Interpolation Method (Case Study; Jam And Riz M.: Academy, 384 . Basin-Assaloyeh). Journal of Applied Sciences, Vol. 8(1), 104-111. Krugman, P., 1991, Geography and Trade. Teslenok, S. A., Nosonov, A. M. and Teslenok, K. Cambridge, MA: MIT Press, 85. S., 2014, Geoinformation Modeling of the Diffusion of Innovations. InterCarto/Intergis- Lurie, I. K., 2008, GIS Mapping. Methods of 20: Sustainable Development of Territories: Geoinformatics and Digital Processing of Cartographic and Geoinformation Support. Satellite Images. Moscow: KDU, 424. Moscow, 159-169. Zamyatina, N., Poturaeva, A., Pelyasov, A. and Makar, S. V. and Nosonov, A. M., 2017, Goncharov, R., 2020, The Sandwich of Russian Assessment and Spatial Regularities of the Space: How Different Spaces Differentiate Innovative Activity Development in the Regions Themes in Regional Science. Regional Science of Russia. Economics, Taxes & law, Vol. 10(4), Policy & Practice. Vol. 12(4), 559-577. 96-106. Wang, X., Lv, J., Wei, C. and Xie, D., 2010, Modeling Spatial Pattern of Precipitation with Manju, K. P., Soraganvi, V. S. and Lingadevaru, D. GIS and Multivariate Geostatistical Methods in C., 2019, Use of RS and GIS for Land Use/Land Chongqing Tobacco Planting Region, China. Cover of Historical Badami Town, Bagalkot International Conference on Computer and District, India. International Journal of Computing Technologies in Agriculture. Innovative Technology and Exploring Springer, Berlin, Heidelberg. 512-524. Engineering (IJITEE), Vol. 8(11) , 2936-2938. Matějíček, L., Engst, P. and Jaňour, Z., 2006, A GIS-Based Approach to Spatio-Temporal Analysis of Environmental Pollution in Urban Areas: A Case Study of Prague's Environment Extended by LIDAR Data. Ecological Modelling, Vol. 199(3), 261-277. Meng, Y., 2020, Spatial Distribution Patterns of Chosen Chemical Elements with Different Spatial Scales in London Topsoils and Lake International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

81 Alcohol Consumption in Spatial Dimension Chaikaew, N., 1, 2* Pimmasarn, S.,3 Prommarin, N.,4 Usiri, P.5 and Sanguansermsri, K.6 1Research Unit of Spatial Innovation Development, School of Information and Communication Technology, University of Phayao, Phayao, Thailand, E-mail: [email protected] 2Geographic Information Science, School of Information and Communication Technology, University of Phayao, Phayao, Thailand 3Remote Sensing and GIS, Asian Institute of Technology, Pathumthani, Thailand 4Political Science, School of Political and Social Science, University of Phayao, Phayao, Thailand 5Science and Technology in Sports, School of Science, University of Phayao, Phayao, Thailand 6School of Education, University of Phayao, Phayao, Thailand *Correspondence Author Abstract The objective of this study is to present the Prevalent Rate of alcohol Consumption in Phayao through the Spatial dimension. The number of a sampling group is 4,830. Those are Phayao population aged between 10 and 70 years old. The study presents the result analysis with the Prevalent Rate, Moran's Index and Kernel Density Estimation. The study finds that the prevalence of alcohol consumption of Phayao population is 54.6 and the Spatial Distribution Pattern of consumption is a random pattern. The alcohol consumption is prevalent in the north and the west of Phayao, especially municipality of Mueang Phayao district, Dok Khamtai district and Chiang Kham district. The output of the study shall be deemed as an academic research. It is to examine the real situation of the alcohol consumption in Phayao province. This can be further used as for decision- making on planning, preventing, controlling and monitoring problem. 1. Introduction Thai populations and 35.4 % or the highest proportion are in northern whereas 32.8 % in northeastern, Alcohol consumption is the crucial health problem in respectively. The highest numbers or 16.0 % of most countries around the world. As per the World Health Organization’s report, it notes that throughout populations drinking more frequently or at least once a week show in northern and ranking in order as the world there have been populations caused to central region at 13.7%. The top 5 of provinces death due to injury and disease attributable to alcohol consumption approximately 3 million ones per year reflecting the alcohol prevalence in older age groups or accounting for 5.3 % of all deaths worldwide consist of Chiang Rai, Lamphun, Phayao, Nan and whereas 7.2 % have early died from disease and illness Surin representing 45.3, 44.1, 44.0, 42.4 and 40.6 %, respectively. According to the afore mentioned data, related to alcohol consumption particularly from injury caused by accidents and suicide (WHO, 2018) the numbers of northern drinkers are proportionately due to excessive alcohol consumption (Demirkol et higher than an overall of the country (National al., 2011). The 2016 report showed that an Statistical Office, 2017).This is in line with the survey approximate 2.3 billion of populations around the world or 43 % of the whole were heavy drinkers with data on local liquor factories of the Center of Alcohol an average/person/year was 6.4 liters of pure alcohol. Studies (CAS) stating that in 2013, there were 3,800 It is forecasted that in future the consumption liquor factories in rural areas and a half of them were quantity, particularly in Southeast Asia, shall located in northern. This therefore indicates increasingly reach to 1.7 liters/person/year in 2025 (WHO, 2018). For Thailand, alcohol prevalence tends behavioral tendency of the alcohol consumption and to be higher every year. As per the survey data of the production which appear apprehensively high (Center of Alcohol Studies, 2014) particularly in Phayao that National Statistical Office of Thailand in 2017, it is found that the numbers of heavy drinkers or alcohol once in 2011 used to be the top of provinces having drinkers are 15.89 million accounting for 28.4 % of all the highest alcohol consumption rate of Thailand and currently its consumption rate remains on top and International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

82 higher continuously. Most populations drinking habit community level shall be deemed as an academic research to examine the real situation in this area to are partly influenced by their northern culture be further used as database of alcohol-related harm in particularly in a rural agglomerative and united parallel with the implementation of task in community society whose lifestyle relies on compliance with the government policies of driving agriculture and cooperation of neighbors or people in health and social problem solving so as to be the same society. According to the study of Wises conducted properly, acceptable, consistent with the target, minimize time spent and budget of effective Sujinnapram and Sahathaya Wises in 2016, it solving problem of alcohol consumption in Phayao. explained that in Phayao communities, alcohol was consumed during both traditional events i.e. wedding 2. Research Objective To analyze the spatial patterns of alcohol ceremony, housewarming, other celebrations and consumption in Phayao province, Upper Northern temple festival (Poi Luang Festival) as well as thank- Thailand, in terms of their geographical distributions. you party for people rendering assistance for crop 3. Research Methodology harvest. Consequently, the local liquor or homemade The Research Methodology is summarized in the flowchart shown in Figure 1. liquor then becomes a part of such social activities and helps economize the expenses of buying other 3.1 Data Acquisition fermented or distillated ones. In addition, as a result 3.1.1 Research papers relevant to the prevalence and alcohol consumption behavior of people in upper of the rapid growth of Phayao, it causes an northern area were studied and reviewed. availability of alcoholic beverages (Sujinnapram et al., 2016) as currently appeared that there are plenty 3.1.2 Population consists of people at age between 10 – 70 years old in Phayao house registration, totally of supplying sources or alcohol selling shops in both 401,271 persons (data gained from statistic system of urban and rural areas which provide more house registration as of December 2017). convenience of availability of alcoholic beverages for both old and new drinkers (Ministry of Public Health, 3.1.3 Sampling group consists of people at age 2015) particularly in pubs near the main educational between 10 – 70 years old in Phayao calculated by prevalence of Phayao drinkers in 2017 (P = 0.44) institutes of the province including Phayao Lake at (National Statistical Office, 2017) with 5% deviation the municipal area, the center of the province and of the prevalent rate (e = 0.022), design effect = 2.5 pubs. gaining the sampling group of 4,889 participants. The random sampling was conducted, without As mentioned above, the alcohol consumption nonprobability sampling, by quota sampling behavior of Phayao populations are diversified and scattering based on population size in each sub- different in each area, both in urban and rural ones, district (68 sub-districts) of Phayao. in view of availability including type or category of alcoholic beverages. Understanding the alcohol 3.1.4 The survey on alcohol consumption behavior was conducted by using a developed questionnaire consumption in spatial dimension and consumption with the sampling group of 4,889 participants from 9 prevalence scattering in each area of Phayao shall districts (68 sub-districts) of Phayao with no evidence thereby contribute to project or activity arrangement record of names, surnames or actual addresses to to promote and launch a campaign to reduce abandon prevent them from an unintentional impact. The and abstain from alcohol consumption that are respondents were not charged for answering the appropriate and consistent with the social contexts of questionnaire, no risk or negative effect that maybe each different area. Currently, the spatial information incurred. Moreover, the research was conducted with the respondents, willingness and in case of and map are considerably applied for public health unwillingness or reluctance during answering, their works since both can indicate location and dispersion withdraws can be made at all times. of health data and spatial analysis to demonstrate size, quantity, dispersion direction, density and difference of spatial health data as well as correlation between prevalence and dispersion of health incidents focusing on characteristic of geographic environment or relevant factors that shall be very helpful for planning the development of health utility and spatial surveillance of people health problem (Matthews, 1990). Thus, the study on alcohol consumption of the people in Phayao through analysis of the prevalent rate, prevalent dispersion pattern and reflection of the prevalent rate or density of the alcohol consumption in spatial dimension at International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

83 All these data were incorporated into a geographic demonstration of spatial information in view of information system (GIS). current consumption behavior (no drinking, rarely drinking (at average of not less than 4 times/month), 3.2 Data Preparation regular weekly drinking (at average of at least once a 3.2.1 The prevalent rate of alcohol consumption was week) and regular daily drinking) whereas those calculated, then dividing the number of drinkers by the number of all samplings and forecasted at 95% information and research findings shall not be confidence interval: CI (Hosiri et al., 2016). disclosed or demonstrated at an individual level but at the sub-district level including an overview of the 3.2.2 The alcohol consumption prevalence of Phayao province. people was analyzed and concluded through Figure 1: Geospatial trend of alcohol consumption International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

84 3.3 Data Analysis ������ℎ������ ������������������������������������������������������ ������������������������������������������ = 1 ������ 3 ������������ (1 − (������������ 2 2 ] 3.3.1 The classic index of spatial autocorrelation ������ 2 ∑[ (Moran’s I) was used to evaluate autocorrelation in the ������ )) spatial distribution of alcohol consumption. Moran’s ������ ������ =1 I was the best way to measure the spatial autocorrelation (feature similarity) based on both Equation 3 feature locations and feature values simultaneously. The spatial statistic was given as: where ������������ was the prevalent rate attribute of sub- district feature ������ within the radius distance (������) of the kernel centered, ������������ was the distance between sub- district feature ������ and the kernel centered. ������������������������������′������ ������ = ������ ∑������������=1 ∑������������=1 ������������,������ ������������������������ The calculated prevalent density was then multiplied ������������ ���������2��� by the number of sub-district feature points, or the sum of the prevalent rate field if one was provided. In Equation 1 this study, the radius of each cone was set to 6,327 where ������������ was the deviation of the prevalent rate meters (an average of nearest distance between sub- attribute for sub-district feature ������ from its mean (������������ − districts) that was estimated to reflect the intensity of ���̅���), ������������,������ was the spatial weight between sub-district alcohol consumption. Each cell on the map surface feature ������ and ������, ������ was equal to the total number of was assigned KDE such that cells at the center of the sub-district features, and was the aggregate of all the cone receive higher prevalent rate estimates, and cell spatial weights: at the cone’s periphery receive smaller prevalent rate estimates (Parzen, 1962, Osei and Duker, 2008 and ������ ������ Chainey and Ratcliffe, 2005). ������������ = ∑ ∑ ������������,������ 4. Results ������=1 ������=������ Among 4,889 respondents, the data of 4,830 were collected accounting for 98.8% as per the following Equation 2 findings: In this study, given a set of sub-district features and 4.1 The analysis result of alcohol consumption an associated the prevalent rate attribute, it evaluates prevalence of the sampling group shows that whether the spatial pattern presented was clustered, currently there are alcohol consumption in 2,640 dispersed, or random. When the index was calculated, respondents representing the prevalent rate of 0.546 the value was between -1 and 1. In case the value was or 54.6% and it is forecasted that in fact, the drinkers will be accounted for approximately 53% to 56 % nearly 1, it indicated that the prevalent dispersion was (Prevalent Rate = 54.6; 95% CI: 53.2, 56.0). Considering the clustered pattern whereas the value nearly -1, it was the dispersed pattern. in case of the value was and classifying based on their consumption behavior, it is found that the consumption of the sampling equivalent to 0, therefore it was the random pattern group is rarely/infrequent drinkers (at average of or variable pattern (Goodchild, 1986, Harries, 1999, lower than 4 times/month) and the number of them is Nakhapakorn and Jirakajohnkool, 2006 and Fang et 1,088 or equivalent to 0.225 prevalent rate or 29.5 % al., 2006). The statistical significance level for this research was set at 0.05. while it is forecasted that in fact, there may be rarely/infrequent drinkers accounted for 3.3.2 Previous spatial autocorrelation analyses approximately 21 to 24 % (Prevalent Rate = 22.5; 95% CI: 21.4, 23.7) whereas 1,423 are the weekly drinkers evaluated the spatial distribution pattern of alcohol (at average of at least once per week) equivalent to consumption only the global level. In the local level, 0.295 prevalent rate or 29.5 % and in fact there may be kernel density estimation (KDE) can be used to detect approximately 28 to 31 % weekly drinkers (Prevalent Rate = 29.5; 95% CI: 28.2, 30.7). The number of the the hotspot of observed events by creating a continuous surface representing the prevalent density respondents who are the daily drinkers or drink every of alcohol consumption per unit area. Across the day is accounted for 129 which is equivalent to 0.027 prevalent rate or 2.7 % and in fact there may be study area, KDE created a set of cones or kernel centered over each sub-district feature points, generating a continuous map of the prevalent density. The predicted density at a new location was determined by the following equation: International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

approximately 2 to 3 % daily drinkers (Prevalent Rate 85 = 2.7; 95% CI: 2.2, 3.1). alcohol consumption is not limit to any specific area 4.2 The analysis result of the prevalent dispersion and can be thereby arisen in all areas of Phayao. pattern of alcohol consumption is as shown in Table 1. Considering based on Moran’s Index, it is found 4.3 The analysis result of the prevalent density of that overall value of the prevalent dispersion of alcohol consumption per unit area can be alcohol consumption in Phayao is nearly 0 meaning demonstrated in a form of spatial information or map that the prevalent dispersion is the random pattern or for more understanding about the prevalence or uncertain pattern; in other word, high prevalence of consumption behavior of the people in each area of Phayao as shown in Figure 2. Figure 2: The current prevalent density of alcohol consumption (A), The prevalent density of rarely/infrequent alcohol consumption (B), The prevalent density of weekly alcohol consumption (C) and The prevalent density daily/everyday alcohol consumption (D) Table 1: Spatial Prevalent Dispersion Pattern of Alcohol Consumption in Phayao Consumption Behavior Moran’s I P-value Spatial Dispersion Pattern Rarely/Infrequent drinking 0.023 0.544 Random Pattern Weekly drinking 0.026 0.502 Random Pattern Daily drinking -0.012 0.967 Random Pattern Currently keep drinking 0.017 0.608 Random Pattern International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

86 Applying Kernel density estimation method to making on planning for appropriate and efficient display the results of the prevalent density of alcohol provision of substantial and adequate resource for consumption per unit area in a form of continuous handling, preventing, controlling and monitoring surface can obviously indicate the different prevalent problem (Kongman, 2010). dispersion of alcohol consumption in each area of Phayao. In other words, the prevalent density of This study concentrated on the issue of analyzing and demonstrating of consumption behavior of alcohol consumption in Phayao currently appears in Phayao people in spatial dimension reflecting an the plain areas between mountains where people overview at the district and provincial levels only reside and highest density is equivalent to 3.505 % per excluding the analysis of significant difference or 1 sq. kilometer. The current prevalent density of correlation between factors affecting alcohol consumption behavior in each area; individual factor consumption also appears in the north and west of the (gender, age, income, occupation, knowledge on province particularly in municipal area of Mueang consumption and alcohol-related harmful side effect, Phayao, Dok Khamtai and Chiang Kham Districts etc.), social context (alcohol consumption of close having urbanization and high population density per persons) and physical & cultural environment unit area. Considering the prevalent rate in terms of (number and location of beverage supply shops, pubs, attitude towards alcohol consumption), etc. (Ministry consumption behavior classified into 3 groups, it is of Public Health, 2015, Kimwatu et al., 2015, found that the current prevalent rate of Papomma, 2015, Hashim et al., 2017 and Changkit rarely/infrequent consumption in Mueang Phayao and Nualchawee, 2018). These factors can be and Dok Khamtai Districts are different from the analyzed, explained and interpreted in spatial prevalence of weekly consumption (at average of at dimension to enhance understanding about alcohol least once per week) whose density of consumption consumption behavior of Phayao people in terms of both direction/dispersion pattern and quantity of appears in the north of the province or in Chiang Kham and Phu Sang Districts and the prevalence of correlation that are varied in each area to be more daily/everyday consumption obviously appears in the completed, reliable and further theoretically referable. commercial zone of Mueang Phayao District close to Phayao Lake and municipal areas of Dok Khamtai Acknowledgements District. This study has been supported by Thai Health Promotion Foundation (ThaiHealth). 5. Discussion and Conclusion References In spatial dimension, Phayao prevalent rate of alcohol consumption is not limit to any specific area and can Center of Alcohol Studies, 2014, The Situation of occur in all areas of both urban and rural ones since Alcohol Consumption and its Impact in Thailand probably its society and areas mostly have similar 2013. Bangkok: Center of Alcohol Studies, Thai community culture of which lifestyle relies on Health Promotion Foundation. agriculture and community activity required for collaboration of people in the same society. The Chaikaew, N., Tripathi, N. K. and Souris, M., 2009, alcoholic beverages are traditionally used to be a part Exploring Spatial Patterns and Hotspots of of activity arrangement for festivals, celebration or Diarrhea in Chiang Mai, Thailand. International other activities within their community (Sujinnapram Journal of Health Geographics, Vol. 8(36), DOI: and Wises, 2016). Demonstration of the prevalent 10.1186/1476-072X-8-36. Chainey, S. and Ratcliffe, J. H., 2005, GIS and Crime data by applying Kernel density estimation method Mapping. England: John Wiley & Sons Ltd. can obviously indicate location, size, dispersion Changkit, N. and Nualchawee, K., 2018, A Study of pattern, density of alcohol consumption data as well as difference of the prevalent rate in each Phayao area Distribution of the Store of Alcoholic Beverage (Chaikaew et al., 2009 and Pimsawan, 2010), help around School by Geographic Information System. Veridian E-Journal, Silpakorn provide a guideline for risk and health impact University, Vol. 11, No. 2 3163-78. assessment attributable to alcohol consumption of the people at area level of authorities concerned with public health of the province. Those authorities can utilize these spatial data showing risk level, incident and severity level including scope of impacted or at- risk areas related to people health problem attributable to alcohol consumption for their decision- International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Demirkol, A., Haber, P. and Conigrave, K., 2011, 87 Problem Drinking - Detection and Assessment in General Practice. Aust Fam Physician, Vol. 40(8), Matthews, S. A., 1990, Epidemiology Using a GIS: 570-574. the Need For Caution. Computers, Environment and Urban Systems, Vol. 14(3), 213-221. Fang, L., Yan, L., Liang, S., de Vlas, S. J., Feng, D., Han, X., Zhao, W., Xu, B., Bian, L., Yang, H., Ministry of Public Health, 2015, Alcohol Behavioral Gong, P., Richardus, J H. and Cao, W., 2006, Modification in Working Age. Bangkok: Health Education Division, Ministry of Public Health. Spatial Analysis of Hemorrhagic Fever with Renal Syndrome in China. BMC Infectious Nakhapakorn, K. and Jirakajohnkool, S., 2006, Diseases, Vol. 6(77), 1-10. Goodchild, M. F., 1986, Spatial Autocorrelation, Temporal and Spatial Autocorrelation Statistics Concept and Techniques in Modern Geography. of Dengue Fever. Dengue Bulletin, Vol. 30, 177– Norwich, United Kingdom: Geo Books. 183. Harries, K., 1999, Mapping Crime: Principle and Practice. Washington DC: Office of Justice National Statistical Office, 2017, The Smoking and Programs. US Department of Justice. Drinking Behaviour Survey 2017. Bangkok: Hashim, S., Lerdsuwansri, R. and Srihera, R., 2017, National Statistical Office. Factor Affecting to Senior High School Alcohol Osei, F. B. and Duker, A. A., 2008, Spatial Drinking in Pathum Thani Province. Thai Journal Dependency of V. Cholera Prevalence on Open of Science and Technology, Vol. 6(1), 1-10. Space Refuse Dumps in Kumasi, Ghana: A Hosiri, T., Sittiun, C. and Limsricharoen, K., 2016, Spatial Statistical Modelling. International Journal of Health Geographics, Vol. 7(62), 1-17. Drinking Behavior and its Prevalence in Grade 10th Students. Journal Psychiatric Association Papomma, S., 2015, Personal Factors and Adverse Thailand, Vol. 61(1), 3-14. Kimwatu, D. M. and Odera, P. A., 2015, Use of GIS Health Behaviors with Drinking and Alcohol Regularly in the One District of the Chaiyaphum Technology in the Implementation of Alcoholic Province. The Office of Disease Prevention and Drinks Control Act: A Case Study of Karima Control 7, Vol. 22(1), 1-8. Location, Othaya, Kenya. International Journal Parzen, E., 1962, On the Estimation of a Probability of Science and Research, Vol. 4(8), 2052-2058. Density and Mode. Annuals of Mathematical Kongman, S., 2010, The Application of Geographic Statistics, Vol. 33, 1065–1076. Pimsawan, N., 2010, The Spatial Analysis of Lung Information Systems for Epidemiological Cancer Patients in Chiang Mai Province. Chiang Surveillance. Academic Journal: Faculty of Mai: Chiang Mai University. Architecture, Khon Kaen University, Vol. 9, 76- Sujinnapram, W. and Wises, S., 2016, Alcohol 89. Feminism and Community Self-Management in Phayao Province. Bangkok: October Print. World Health Organization, 2018, Global Status Report on Alcohol and Health 2018. Geneva: World Health Organization. International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Accessibility Analysis of Roads Network in Ma'an Governorate Ayed Taran, A. M. Applied Geography Department, Al al-bayt university, Jordan, E-mail: [email protected] Abstract This study aims at analyzing the level of accessibility of the roads network of different types in Ma'an governorate in the south of Jordan through a set of standards proposed by Shimbel in 1953. The study adopted the descriptive analytical approach in the analysis of geographical information and data related to road networks and urban communities. To achieve the objectives of the study, statistical methods were also used in measuring the accessibility to the road network through some methods such as Shimbel Index and the Associated Number and others using the matrix method. The study found that there are differences in the accessibility to the nodes located on the road network in Ma'an governorate. Ma'an node was the easiest to access because of its central location compared to the other nodes, while Al-Mudawwara, Ras Al-Naqab and Husseiniya nodes were the most difficult to access as a result of their marginal and extreme locations far from the other nodes. 1. Introduction to the optimal utilization of human and natural The change in the accessibility to the road network resources such as phosphate, cement and other plays a positive role in economic and social minerals in southern Jordan, which contributed to development. The ability to transport goods and support the national economy and improve the services is a crucial element for development. standard of living of the inhabitants of those areas. Accessibility also determines the spatial characteristic of a region in relation to other regions; Ma'an governorate is the largest governorate in thus, it contributes effectively to regional Jordan. It is characterized by the widespread urban development (Al-Hussein, 2011). Moreover, communities and the long distances between those accessibility is a key element in the geography of communities. As a result, in habitants face inter- transport in particular and in geography in general. governorate and intra-governorate accessibility The relationship between transport and geography is problems. These problems include the long time of strong, not only because it is related to the branch of arrival and the high cost of transportation, as well as economic geography, but also because transport the numerous means of transport they use during affects other branches of geography such as their journey. Hence, an analysis of the accessibility geopolitics, urban geography and the geography of of the road network in this governorate is required services (Thatcher, 1958). in order to identify the nodes or the central vertices that are easier to reach as well as those which are Accessibility to nodes is an important indicator more difficult to reach. This study aims at achieving that helps to determine spatial relationships within the following objectives: any region, as it is essential in land use planning and transport decision-making. It is also an easy tool to 1. Evaluating the actual accessibility between guide planning and development decisions (Gregory the nodes or vertices in Ma'an Governorate et al., 2009). Accessibility has become a prerequisite and classifying the urban communities for achieving integration of urban areas centers and according to the actual accessibility. the surrounding areas, as well as for organizing the spatial area of entire cities and urban centers (Li and 2. Identifying the nodes or vertices that have Lu., 2005).The development of the transportation high accessibility, thus forming central sector in Jordan is a key factor in the process of vertices in Ma'an Governorate. economic development. The establishment of road networks has helped to link areas to each other and 3. Identifying the nodes or vertices that have thus facilitate getting around among them. It also reduced accessibility, thus form in helped in the creation of the so-called regional peripheral or marginal vertices in Ma'an specialization in agriculture and in industry. Governorate. Furthermore, the establishment of road networks led International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

4. Determining the status of the road network 6. Vandenbulcke et al. (2009) a study on and its role in linking urban communities accessibility analysis in Belgium for in Ma'an Governorate. transport planning and land-use. The study addressed the definition of accessibility 2.Previous Studies and its multiple measures by building a Several studies have dealt with the issue of matrix of origin and destination based on accessibility in various places. These studies have the distance and time of the journey. applied many statistical methods in this regard. These studies include the following: 7. Ghallab (2014) a study on the geographical assessment of the spatial accessibility to 1. Aldagheiri (2014) a study on the analysis health services in Kafr El Dawwar of accessibility of the roads network in Al- countryside in Egypt using the geographic Qassim region in Saudi Arabia by information systems. The study concluded calculating the number of direct that some towns such as Zahra, King connections between urban centers and the Osman and Abis Al-Mustagida are the number of nodes between each two major easiest to reach the Central Hospital in nodes. The study concluded that the road Kafr El Dawwar while the two towns of the network in the region is capable of Tarh and Sidi Ghazi are the most difficult contributing to economic development and to reach. reducing distances between major urban centers. 3. Study Area Ma'an Governorate is located in the southern region 2. Wazi (2012) a study on accessibility of Jordan, extending between longitudes assessment in Sana'a city using isochron 35°and38°E, and latitudes 29°12 -31°12N. The maps. The study relied on time to measure administrative boundary reaches north to the border the accessibility between the city center of Amman Governorate. The eastern and southern and its outskirts in an attempt to identify borders of Ma'an are the border between the the causes of low accessibility in the city Hashemite Kingdom of Jordan And Saudi Arabia. and then develop a plan to address these The study area also has borders with Karak and causes. Tafila governorates in the north and west and the governorate of Aqaba in the west, as shown in 3. Al-Dosari(2011) a study on the traffic jams Figure 1. The study area is a link between the most and the accessibility to Kuwait City. The important governorates of the Kingdom (Al- study relied on the equal-time lines on the Fanatsah, 2015). axes of road network from the outskirts of the city towards its center. The study Ma'an Governorate is the largest in the identified the most important connections Hashemite Kingdom of Jordan. It covers an area of that have accessibility problems and 32832 km2, which is 37% of the area of the suggested solutions. Kingdom (Department of Statistics, 2015). Ma'an governorate is administratively divided into four 4. Ighraieb (2010) a study on assessing the districts, as shown in Figure 2, namely: Ma'an (Al- accessibility to urban nodes on the road Qasaba) district, which includes about 34 urban network and their degree of centrality in communities; Petra district, 12 urban communities; Hebron Governorate. The study adopted Shobak district, 14 urban communities; and Al- the method of matrices and their indicators Husseiniya district with three urban in order to determine the centralization of communities.According to the General Population the nodes. The study showed that Hebron and Housing Census, the population of Ma'an was ranked first in accessibility according governorate reached about 148000 in 2015, with a to the criteria used in the study. population density of (4.5 per km2) of the total population density of the Kingdom (57.4 per 5. Scheurer and Curtis (2007) on accessibility km2).The governorate's population is divided over measures: overview and practical the four districts mentioned above as follows: Ma'an applications. The study discusses seven (Al-Qasaba) district (89.497), Petra district categories of accessibility measures: spatial (20.380), Shobak district (19.820), and Al- separation model, contour measures, Husseiniya district (17.810), (Department of gravity measures, competition measures, Statistics, 2015). time-space measures, utility measures, and network measures; as well as the rules that govern each of these categories. International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

urban communities were analyzed. Some methods were used to achieve the objectives of the study, including utilizing statistical methods in measuring the accessibility of the road network through usingShimbel Index and the Associated Number as well as using the matrix method. Furthermore, geographic information system (GIS), especially the ArcGis 10.2 software was used to create the topological map of the road network which converts the complex road network to a simple abstract network that consists of lines and points, neglecting the direction and the real distance of the roads and traffic density on them. Figure 1: Study area 4.2 Measuring Accessibility There are multiple classifications and measurement Figure 2:Districts of Ma'angovernorate methods for accessibility. For example, the 4.1 The Descriptive Analytical Approach classification of Handy and Niemeier in 2007 which The study adopted the descriptive analytical classified accessibility into three categories, namely approach. The locations of urban communities were the Isochrone scale, an interaction-based scale, and examined in terms of their distribution and a utility-based scale. Another classification expansion in Ma'an governorate. Geographical suggested four basic accessibility criteria: an information and data related to road network and infrastructure-based scale, an activity-based scale, a passenger-based scale, and a utility-based scale (Gregory et al., 2009). The classification on which the study was based is the one proposed by Shimbel (Rodrigue, 2020), based on the construction of a matrix to measure accessibility and communication between the nodes of the network (Abu Radhi, 1989). Accessibility is relative, depending on a number of geographic factors that distinguish nodes and the connections that link them to each other. Thus, preparing an Accessibility Matrix is one of the best quantitative methods to measure the relative importance of the nodes on transportation routes (Al-Zouka, 2005). The matrix is an appropriate means of storing data and showing the distances in the transport network. It is a table placed on the axes of the nodes or studied vertices to clarify the relations between them. The size of this relationship varies according to the various variables used in measuring accessibility such as the number of connections ending at each node and the distance between the nodes (Abdo, 2007). The value of zero is given for direct connection that does not need to be changed, while values are increased by the increase in the number of connections between the nodes or vertices in the network; and decreases when the connection between one node and the other is easy (Abu Assi, 2011). 4. Results and Discussion Accessibility reflects the ease of movement within the region or territory. This ease reflects the comprehensiveness of the road network and its International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

ability to connect areas to each other. The more contains a large number of roads and nodes, as in roads (direct connections) between nodes or the case of Ma'an governorate, it is difficult to vertices, the easier the transition from one node to identify easily accessible nodes and inaccessible another. Thus, accessibility to these nodes is nodes. In that case we turn to other methods to increased (Muhammad, 2003). The accessibility measure accessibility (Muhammad, 2003). indicator is one of the most important quantitative indicators used to measure the ease of access to any Any road network has to be simplified and node or station in the network. The node on the road converted to a topological map prior to being network is described as accessible or inaccessible. studied and analyzed (Abu Hajjaj, 1989). Therefore, An accessible node is a node that is located near the road network in Ma'an governorate has been other nodes on the network and can be accessed converted into a topological form using the ArcGIS with as few roads as possible (Al-Khashman, 2013). 10.1 program. Due to the large number of populated areas in the governorate; the study was limited to the selection of the largest and most important populated places, which included 19 communities to cover the study area as shown in Figure 3. The road network in Figure 3 consists of nodes (peripheral populated areas), vertices (central populated areas), and edges (the roads that link nodes and vertices together). 4.1 Accessibility According to the Number of Connections between the Nodes (Shimbel Index) To calculate accessibility according to the number of connections between nodes, we determine the number of connections in the matrix. Then, the nodes are arranged according to accessibility on the basis that the node that is connected to the rest of the other nodes through the least number of connections is the most accessible; while the node that is connected to the rest of the other nodes through the largest number of connections is the least accessible as shown in (Table 1). By applying this variable on the road network in the study area as shown in (Figure 4) we find the following: Figure 3:Topological Map of Ma'angovernorate 1. There isn't any node or vertex in the road network that connects directly to all other Determining the accessibility to the nodes or nodes or vertices. vertices located on the road network is beneficial for regional and urban planning processes. It guides 2. Ma'an node is ranked first in terms of decision-makers in selecting the optimal location for accessibility with the lowest number of government services and facilities such as hospitals, connections (35); thus, it is a central node. universities, schools, civil defense centers, etc. Eil node came second with (39) connections, These facilities are usually located in the most then came Al-Tahounah node in third place central urban centers for ease of access from all with (40) connections. parts of the region or city in the shortest and least number of routes, thus reducing the distance and 3. Some of the other nodes came in the same thereby reducing journey time and cost (Al- rank: two nodes came in the fourth place Khashman., 2013).It is so easy to identify nodes or with (42)connections, two nodes came in the central and remote vertices if the road network is fifth place with (43) connections, two nodes simple; but when the road network is complex and came in ninth place with (50) connections, two in the tenth place with (52) connections, and two in the thirteenth place with (56) connections. 4. Al-Husseiniya node came in the last place with the highest number of connections (58). This is due to the fact that it occupies a peripheral and marginal position in the International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

network and is isolated from the rest of the longitudinal extension of the road axes in nodes in the far north of the governorate. Ma'an Governorate from north to south. 5. The distribution of some of the network nodes that are close to each other reveals the Figure 4: Nodes ranks according (Shimbel Index) Table 1: Accessibility Matrix according to the number of connections between nodes (Shimbel Index) Nodes Jafr Mudawwara RasNaqab Mraigha Sadaqa Rajef Taybeh Eil Bir Abu Danneh Bastah Tahouneh WadiMousa Rashid Uthruh Shobak Almothaleth Husseiniya Hashemiya Ma'an Ma'an - 1 2 232 3 3 1 2 3 1 2 3 2 1 2 1 1 Hashemiya 1 - 1 122 3 3 2 3 4 2 3 4 3 2 3 2 2 Husseiniya 2 1 - 233 4 4 3 4 5 3 4 5 4 3 4 3 1 Almothaleth 2 1 2 - 11 2 2 2 3 3 4 3 4 5 4 5 3 3 Shobak 3 2 3 1-2 2 1 3 2 3 3 2 3 4 5 4 5 4 Uthruh 2 2 3 12- 1 2 1 2 2 3 3 4 4 3 4 3 3 Rashid 3 3 4 221 - 1 2 2 1 3 2 3 4 4 4 4 4 WadiMousa 3 3 4 212 1 - 2 1 2 2 1 2 3 4 3 4 4 Tahouneh 1 2 3 231 2 2 - 1 2 2 3 4 3 2 3 2 2 Bastah 2 3 4 322 2 1 1 - 112 3 2 3 4 33 Bir Abu Danneh 3 4 5 3 3 2 1 2 2 1 - 2 3 4 3 4 5 4 4 Eil 1 2 3 4 3 3 3 2 2 1 2 - 1 2 1 2 3 2 2 Taybeh 2 3 4 323 2 1 3 2 3 1 - 1 2 3 2 3 3 Rajef 3 4 5 434 3 2 4 3 4 2 1 - 3 2 1 4 4 Sadaqa 2 3 4 544 4 3 3 2 3 1 2 3 - 1 2 3 3 Mraigha 1 2 3 453 4 4 2 3 4 2 3 2 1 - 1 2 2 RasNaqab 2 3 4 544 4 3 3 4 5 3 2 1 2 1 - 3 3 Mudawwara 1 2 3 353 4 4 2 3 423 4 3 23 -2 Jafr 1 2 1 343 4 4 2 3 4 2 3 4 3 2 3 2 - Total 35 43 58 50 52 45 49 44 40 42 55 39 43 56 52 2 56 54 50 Rank 1 5 14 9 10 7 8 6 3 4 12 2 5 13 10 4 13 11 9 International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

4.2 Accessibility According to the Associated Number, as the values of the associated number Number ranged between 3-5 connections. The associated number indicates the highest number 2. Ma'an node was ranked first in terms of in the cells of the column or row connected to the accessibility according to the associated number vertex or node in the accessibility matrix. To and thus represents a central node. It had the calculate accessibility according to the value of the lowest value of the number associated to it at 3 number associated to the nodes, we determine the connections, which indicates that it has a central highest value connected to each vertex in the matrix. location. Then, the nodes are arranged on the basis that the 3. Many nodes in the study area had the same rank node which is connected to the other nodes with the of accessibility. Nine nodes had the second rank lowest number associated to the nodes is the most and other nine had the third rank in accessibility. accessible to the rest of the nodes of the network, 4. It was found that Husseiniya node or vertex while the node that is connected to the rest of the came in last place in accessibility. This nodes with the largest number associated to the corresponds to the result of the analysis of nodes is the least accessible (Table 2). When accessibility according to the number of applying this variable to the road network in the connections between the nodes. study area, as shown in (Figure 5), we find the 5. Most of the nodes that were inaccessible following: according to the number of connections between the nodes were ranked last in accessibility 1. There is no significant difference between the according to the Associated Number, such nodes or vertices of the study area in RasNaqab, Rajef, Abu Danneh, Mudawwara, accessibility according to the Associated Shobak and others. Table 2: Accessibility Matrix according to the number of connections between nodes (Associated Number) Nodes Jafr Mudawwara RasNaqab Mraigha Sadaqa Rajef Taybeh Eil Bir Abu Danneh Bastah Tahouneh WadiMousa Rashid Uthruh Shobak Almothaleth Husseiniya Hashemiya Ma'an Ma'an - 1 2 2 3 2 3 3 1 2 3 1 2 3 2 12 1 1 Hashemiya 1 - 1 1 2 2 3 3 2 3 4 2 3 4 3 23 2 2 Husseiniya 2 1 - 2 3 3 4 4 3 4 5 3 4 5 4 34 3 1 Almothaleth 2 1 2 - 1 1 2 2 2 3 3 4 3 4 5 45 3 3 Shobak 3 2 3 1 - 2 2 1 3 2 3 3 2 3 4 54 5 4 Uthruh 2 2 3 1 2 - 1 2 1 2 2 3 3 4 4 34 3 3 Rashid 3 3 4 2 2 1 - 1 2 2 1 3 2 3 4 44 4 4 WadiMousa 3 3 4 2 1 2 1 - 2 1 2 2 1 2 3 43 4 4 Tahouneh 1 2 3 2 3 1 2 2 - 1 2 2 3 4 3 23 2 2 Bastah 2 3 4 3 2 2 2 1 1 - 1 1 2 3 2 34 3 3 Bir Abu Danneh 3 4 5 3 3 2 1 2 2 1 - 2 3 4 3 45 4 4 Eil 1 2 3 4 3 3 3 2 2 1 2 - 1 2 1 2 3 2 2 Taybeh 2 3 4 3 2 3 2 1 3 2 3 1 - 1 2 32 3 3 Rajef 3 4 5 4 3 4 3 2 4 3 4 2 1 - 3 21 4 4 Sadaqa 2 3 4 5 4 4 4 3 3 2 3 1 2 3 - 12 3 3 Mraigha 1 2 3 4 5 3 4 4 2 3 4 2 3 2 1-1 2 2 RasNaqab 2 3 4 5 4 4 4 3 3 4 5 3 2 1 2 1- 3 3 Mudawwara 1 2 3 3 5 3 4 4 2 3 4 2 3 4 3 23 - 2 Jafr 1 2 1 3 4 3 4 4 2 3 4 2 3 4 3 2 3 2 - Associated Number 3 4 5 5 5 4 4 4 4 4 5 4 4 5 5 5 5 5 4 Rank 1 2 3 3 3 2 2 2 2 2 3 2 2 3 3 33 3 2 International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Figure 5: Nodes Ranks according to the Associated Number Table 3: Accessibility Matrix according to the length of connections between the nodes/km Nodes Jafr Mudawwara RasNaqab Mraigha Sadaqa Rajef Taybeh Eil Bir Abu Danneh Bastah Tahouneh WadiMousa Rashid Uthruh Shobak Almothaleth Husseiniya Hashemiya Ma'an Ma'an - 39 47 63 67 23 33 36 4 23 27 21 30 40 34 26 41 125 66 Hashemiya Husseiniya 39 - 8 29 33 65 72 75 43 62 36 60 69 79 81 65 80 164 60 Almothaleth 47 8 - 37 41 73 80 83 51 70 44 68 77 87 89 73 88 172 54 Shobak Uthruh 63 29 37 - 4 36 46 39 55 53 49 56 49 59 64 89 104 188 87 Rashid WadiMousa 67 33 41 4 - 38 50 35 59 57 53 60 45 55 68 93 108 192 91 Tahouneh Bastah 23 65 73 36 38 - 10 13 19 17 13 20 33 43 28 46 61 148 91 Bir Abu Danneh 33 72 80 46 50 10 - 13 29 7 3 10 23 33 18 36 51 158 99 Eil 36 75 83 39 35 13 13 - 32 15 11 18 10 20 26 41 51 161 102 Taybeh Rajef 4 43 51 55 59 19 29 32 - 19 23 22 31 41 30 34 59 129 95 Sadaqa Mraigha 23 62 70 53 57 17 7 15 19 - 4 3 12 22 11 25 48 148 89 RasNaqab Mudawwara 27 36 44 49 53 13 3 11 23 4 - 7 16 26 15 29 44 152 93 21 60 68 56 60 20 10 18 22 3 7 - 9 19 8 22 37 146 87 Jafr 30 69 77 49 45 33 23 10 31 12 16 9 - 10 18 31 41 155 96 40 79 87 59 55 43 33 20 41 22 26 19 10 - 13 27 31 165 106 34 81 89 64 68 28 18 26 30 11 15 8 18 13 - 14 29 159 100 26 65 73 89 93 46 36 41 34 25 29 22 31 27 14 - 15 151 92 41 80 88 104 108 61 51 51 59 48 44 37 41 31 29 15 - 166 107 125 164 172 188 192 148 158 161 129 148 152 146 155 165 159 151 166 - 191 66 60 54 87 91 91 99 102 95 89 93 87 96 106 100 92 107 191 - Total 1706 2870 1161 909 805 876 755 673 645 685 775 781 771 777 1149 1107 1242 1120 745 Rank 4 14 17 13 15 8 6 9 7 3 1 2 5 11 10 12 16 19 18 International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Figure 6: Nodes Ranks according to the length of connections between them 4.3 Accessibility According to the Length of distance of (151 km.). This is due to its Connections between the Nodes (Distance) location on the road network having no direct The node that is connected to the rest of the network connections with the rest of the other nodes. nodes via the least total of connections lengths is the 4. It can be noted that the importance of the most accessible to the rest of the network nodes. nodes according to the distance between The general rule states that travelers prefer taking them in accessibility is not consistent with the shortest routes to the rest of the network nodes the population, urban and economic (Muhammad, 2003).Based on the number of nodes importance. Ma'an node came in the fourth in the study area, the variable used requires the place, although it represents the population creation of a matrix showing the distances between and economic weight in the study area. This nodes, as shown in Table 3. Based on the is due to the large spatial area of the accessibility matrix according to the length of governorate, which is reflected in the lengths connections between the nodes, we found the of the distances between the nodes. following (Figure 6): 4.4 Accessibility According to the Number of Nodes 1. The connections between the nodes in Ma'an between Each Two Nodes governorate are characterized by being long, This variable is based on the assumption that the which confirms the wide spread of the nodes most accessible node is the one that is connected or urban clusters in the region due to the directly to other nodes without the need to change large area occupied by Ma'an governorate stations. (Al-Ruwaythi, 1992). On this basis, the compared to other Jordanian governorates. matrix can be configured to determine points of change (inter-nodes) between every two nodes on 2. The lengths of the connections between the the road network. Thus, the node that registers the nodes range between 645-2870km. The node lowest number of inter-nodes is the most accessible, of Bir Abu Danneh (645 km) ranked first in as shown in Table 4.We conclude the following accessibility based on the lengths of from the Accessibility Matrix according to the connections between the nodes (distance) number of nodes between every two nodes (Figure with an average distance of about (34 km), 7): followed by Eil, Bastah and Ma'an nodes which ranked second, third and fourth (673 1. The number of inter-nodes between each two km, 685 km, 745 km), respectively. nodes ranges between34 to 59 nodes. The difference between these two values 3. Mudawwara node (2870km) ranked last in represents the intermediate node in its accessibility based on the lengths of connections between nodes with an average International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

position with reference to the road axes in the 4.5 Overall Accessibility of the Road Network study area. This variable is based on the combination of two or 2. Ma'an node is ranked first (34 nodes) in more variables in order to eliminate the defects of terms of accessibility according to the each variable. It can be measured by combining two number of inter-nodes, indicating its central related variables such as the number of inter-nodes position compared to other nodes in the and the length of the connections between the region. nodes. To combine the two variables, we assume 3. Husseiniya node ranked last (59 nodes) in that each change in pathway from one node to terms of accessibility according to the another is equivalent in cost and effort to an average number of inter-nodes. This result of 10 km (Muhammad, 2003). Table 5 shows the corresponds to the result of accessibility overall accessibility between network according to the number of connections nodes.Although the overall accessibility variable is between the nodes, which confirms the based on the lengths of the connections variable and relatively distant location of this node from the number of inter-nodes variable, it is an the rest of the other nodes. important indicator of accessibility between the road network nodes in Ma'an Governorate. Table 4: Accessibility Matrix according to the number of nodes between every two nodes Nodes Jafr Mudawwara RasNaqab Mraigha Sadaqa Rajef Taybeh Eil Bir Abu Danneh Bastah Tahouneh WadiMousa Rashid Uthruh Shobak Almothaleth Husseiniya Hashemiya Ma'an Ma'an - 12 2 2 2 3 3 1 2 31 2 3 2 1 2 1 1 Hashemiya 1 -1 1 2 2 3 3 2 3 42 3 4 3 2 3 2 2 Husseiniya 21- 23 3 4 4 3 4 53 4 54 3 4 3 2 Almothaleth 2 1 2 - 1 1 2 2 2 3 3 3 3 4 5 4 5 3 3 Shobak 2 23 1 - 2 2 1 3 2 33 2 3 4 5 4 4 4 Uthruh 2 23 12 - 1 2 1 2 23 3 443433 Rashid 3 34 2 2 1 - 1 2 4 13 2 3 4 4 4 4 4 WadiMousa 3 3 4 2 1 2 1 - 3 1 2 2 1 2 3 4 3 4 5 Tahouneh 1 23 2 3 1 2 3 - 3 22 3 4 3 2 3 2 2 Bastah 2 34 3 2 2 4 1 3 - 11 2 3 2 3 4 3 3 Bir Abu Danneh 3 4 5 3 3 2 1 2 2 1 - 2 3 4 3 4 5 4 4 Eil 1 2 3 3 3 3 3 2 2 1 2 - 1 2 1 2 3 2 2 Taybeh 2 34 3 2 3 2 1 3 2 31 - 1 2 3 2 3 3 Rajef 3 45 4 3 4 3 2 4 3 42 1 - 3 2 1 4 4 Sadaqa 2 34 5 4 4 4 3 3 2 31 2 3 - 1 2 3 3 Mraigha 1 23 4 5 3 4 4 2 3 42 3 2 1 - 1 2 2 RasNaqab 2 34 5 4 4 4 3 3 4 53 2 1 2 1 - 3 3 Mudawwara 1 2 3 3 4 3 4 4 2 3 4 2 3 4 3 2 3 - 2 Jafr 1 2 2 3 4 3 4 5 2 3 4 2 3 4 3 2 3 2 - Total 34 43 59 49 50 45 51 46 43 46 55 38 43 56 52 48 56 52 52 Rank 1 3 13 7 8 4 9 5 3 5 11 2 3 12 10 6 12 10 10 International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Figure 8 shows the following: the network. Bastah and Taybeh also came in third and fourth places respectively. 1. Eil node preserved the first place of 3. It is noted that there is a slight difference in accessibility to the rest of the network nodes the levels of accessibility of the network with an index of (1.053 km), which confirms nodes according to this variable compared to its central location and ease of the other variables. communication with other nodes in the 4. Jafr and Mudawwara nodes retained the 18th network. and 19th places with (2.226 km) and (3390 km) respectively. These nodes are therefore 2. It was found that Ma'an node also retained having the lowest accessibility, indicating the second position with an index of (1.085 that they have marginal locations that are far km), which indicates that it has an away from the other nodes of the network. intermediate position between the nodes of Figure 7: Nodes Ranks according to the number of nodes between every two nodes Table 5: Overall accessibility of the road network Variables in km. Accessibility No. of nodes X Nodes Actual distance Total (1 + 2) Rank 10 (2) Ma'an (1) 1085 2 Hashemiya 34 x 10 = 340 745 1550 13 Husseiniya 43 x 10 = 430 1120 1832 17 Almothaleth 59 x 10 = 590 1242 1597 14 49 x 10 = 490 1107 1649 15 Shobak 50 x 10 = 500 1149 1227 7 Uthruh 45 x 10 = 450 777 1281 9 Rashid 51 x 10 = 510 771 1241 8 WadiMousa 46 x 10 = 460 781 1205 6 Tahouneh 43 x 10 = 430 775 1145 3 Bastah 46 x 10 = 460 685 1195 3 Bir Abu Danneh 55 x 10 = 550 645 1053 1 38 x 10 = 380 673 1185 4 Eil 43 x 10 = 430 755 1436 12 Taybeh 56 x 10 = 560 876 1325 10 Rajef 52 x 10 = 520 805 1389 11 Sadaqa 48 x 10 = 480 909 1721 16 Mraigha 56 x 10 = 560 1161 3390 19 RasNaqab 52 x 10 = 520 2870 2226 18 Mudawwara 52 x 10 = 520 1706 Jafr International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International

Figure 8: Nodes Ranks according Overall accessibility Finally, based on the results of the analysis, it was result of its marginal location in the far south-east of found that Ma'an city node obtained the first place Ma'an governorate, the same goes for Al-Jafr and according to most of the scales and variables that Ras Al-Naqab nodes. The other nodes ranks varied were relied upon in this study. On the other hand, according to the variable used but to a limited the other nodes' accessibility levels varied according extent, so that some nodes maintained the same rank to the variable used, but this variation was to a in more than one variable or index such as Eil, limited extent as some nodes have maintained the Sadaqa and Rashid nodes. The overall accessibility same rank in more than a variable. The overall variable is a true average of the node's ranks in accessibility variable is a true average of the nodes accessibility as it is based on more than one 'ranks of accessibility as it depends on more than variable. one variable and thus can be relied upon in the overall ranking of accessibility. 6. Recommendations Considering the findings of the study of the 5. Conclusions accessibility to the road network in Ma'an The construction of matrices in geographical Governorate, the Planning for new alternative and research and studies is an important means to obtain shortcut roads that will increase accessibility to the precise and accurate results, especially in the nodes that is difficult to access such as Al-Jafr, analysis of road networks. Ma'an node retained its Mudawwara, Husseiniya and Ras Al Naqab.And position in the first place in accessibility according Working on achieving a spatial balance between the to almost all variables and scales adopted in the levels of accessibility of the accessible nodes as well study, due to its central position from other nodes in as the inaccessible nodes. And Increasing the the network. Ma'an node ranked fourth in attention to the importance of accessibility to the accessibility according to the lengths of the city of Petra in Wadi Musa; as it recorded medium connections variable (distance) with a total of 745 levels of accessibility according to the variables km, due to the wide spread of the nodes in Ma'an used in the study, given its touristic and economic governorate and the vast area of the governorate. importance in the region. and Establishing an Al-Mudawwara, Al-Jafr, Ras Al-Naqab and Al- integrated road database in Ma'an governorate to Husseiniya nodes suffer from difficulty in address accessibility in a scientific way away from accessibility, as these nodes occupied the last places guessing and speculating. And Paying more of accessibility according to the variables used, due attention to the use of modern technologies such as to their marginal location on the road network. geographic information systems GIS by government In addition to that, Husseiniya node ranked last institutions; and the use of road network analysis in according to the number of connections between the the studies of accessibility due to the precise and nodes and the number of inter-nodes due to its accurate results provided by these technologies. peripheral location in the far north of Ma'an Finally, The expansion of such studies and Governorate. Mudawwara node ranked last generalizing this study to other governorates; and according to the length of connections between the conducting comparison between these governorates nodes and the overall accessibility variable as a International Journal of Geoinformatics, Volume 16, No. 4, October - December 2020 Online ISSN 2673-0014/ © Geoinformatics International