جغرافیا و توسعه (Jun 2012)
The Assessment of Geostatistic Methods and Linear Regression in Order to Specify the Spatial Distribution of Annual Precipitation Case study: Boushehr Province Dr. Gholam Ali Mozaffari Assistant Professor of Geography University of Yazd Dr. Seyed Hossein Mirmusavi Assistant Professor of Geography University of Zanjan Younes Khosravi M.Sc Student of Climatology
Abstract
Introduction The time variability of precipitation is considered as a key factor affecting on the structure and functioning of ecosystems, but from the view point of size and scale is far less than the spatial variability (Knapp and Smith, 2001; Austin et al, 2004; Collins et al, 2008). Determine the most appropriate interpolation methods at a regional level and its spatial and location distribution, is necessary for spatial distribution of rainfall. There are different methods to estimate parameters that as the classical methods, such as Thissen and arithmetic average proposed. Although all of these calculations are quick and easy, but for reasons including failure to consider the location, layout and relationship between them, are not of good accuracy. There are other methods that consider the spatial correlation structures of the data are of great importance that such method is geostatistic. In geostatistic, first the presence or absence of spatial structure of the data is presented and then if there is spatial structure, data analysis is performed. Precipitation behavior in each region varies according to altitude. This behavior is described in the regression relationship in relation with height or the distance. On the other hand, because each region has its own spatial characteristics, so follows a certain interpolation method and the results of one region can not be attributed to another region. The aim of this study is to review the relationship between precipitation and elevation based on digital elevation model and then evaluating its results with ordinary kriging and simple models in order interpolate the annual rainfall in Bushehr province. Research Methodology The study area is Bushehr province. from the number of 101 stations in the region, due to short-term period and the selection of suitable sites with good dispersion, only data from 57 precipitation stations with 11-year period (1997-2007) were used. The statistical methods used in this study are as follows: A- Geostatistic methods: Method used for interpolation, is kriging which is the best linear unbiased estimate .. The Assessment of Geostatistic Methods and Linear Regression in ... B - Variogram analysis The main purpose of calculating the variogram is that be able to recognize variability of the variable regard to the spatial or time distance. For performing this, it is necessary to calculate the sum square differences between couples placed at the distance of h from each other and be plotted against h. C- Methods and evaluation criteria Different interpolation method based on the Cross-Validation procedure will be evaluated. In this method, a point is removed temporarily and by using the considered interpolation, a value is estimated for that point. Then the removed value is returned to its place and for the rest of points, this estimate is done separately. D - Linear regression Regression analysis provides the possibility to predict the changes of dependent variables through independent variables and determine the share of each independent variables in explaining of the dependent variable. Discussion and Results A- Analysis of kriging interpolation model in precipitation interpolation Semi variogram was used for spatial analysis of data. For making the best interpolation, the most important step, is presenting an appropriate model of Semi variogram. The models used in this study include: spherical model, exponential model, Gaussian model, circular model, rational quadratic, Tetra spherical and Penta spherical modal which have made with two techniques of simple kriging and ordinary kriging. The best model which is able to explain the spatial distribution of rainfall is the exponential model of ordinary kriging. So with great confidence we can use this model for estimation of rainfall and other parameters used in the region. B- Evaluation of linear regression based on digital elevation model for the interpolation of precipitation There are wide equations for performing interpolation by regression analysis, which selecting the appropriate equation, depends on the correlation value between the secondary and primary variables. For this purpose firstly, the data of rainfall and altitude of the under study were called in ver1.4 Curve Expert software environment by using linear regression models. Then the considered data were fitted by 18 models. Correlation between topography and spatial interpolation methods indicates that the highest correlation exists respectively, in the fourth degree polynomial, quadratic functions and ordinary kriging model with exponential model. Correlation with the topography of the exponential model showed a positive relationship between amounts of precipitation with altitude but this relationship is weaker than the other two methods that this kind of relationship clears the relationship of rainfall and rainfall in the rainfall interpolation. Conclusion 1-The best method for interpolation of annual rainfall in Bushehr province, the fourth degree polynomial regression function was diagnosed. 2 -The use of linear regression methods and using it in the digital elevation model of the earth, shows better the precipitation behavior in the areas where are faced with a lack or deficiency of stations, which itself shows better the value of this approach in environmental studies. 3-One of the principles of kriging interpolation method is the existing of basic point data ,containing a point value to a parameter that the proper and adequate attention to the distribution manner of meteorology stations reduces the errors and increases the accuracy of interpolation. 4-Cluster analysis can be used to verify the homogeneity of selected data in different zones. Because cluster analysis creates the possibility that similar data can be placed in one group and while the data within a group have many similarities with each other, but other groups have significant differences. 5-The final goal of the spatial variation of rainfall is the safe simulation of precipitation data changes in location, so that the next targets, including short-term and long-term forecasts of rainfall in each region is provided. Keywords: Geostatistics, Linear Regression, Digital Elevational Model (DEM), Geographic Information System (GIS), Precipitation, Boushehr province. Refrences 1- Asakereh, H (1387). Application of the kriging interpolation of precipitation (Case study: 12/26/1376 interpolation of precipitation in Iran), Geography and Development, No. 12 2- Austin, A. T., L. Yahdjian, J. M. Stark, J. Belnap, A. Porporato, U. Norton, D. A. Ravetta, and S. M. Schaeffer (2004). Water pulses and biogeochemical cycles in arid and semiarid ecosystems. Oecologia 141. 3- Bndrabady Rahimi, S.Mehdian, MH (1382). Spatial variation of monthly rainfall in arid and semiarid regions of South East Iran, the third regional conference and the first national conference on climate change, Isfahan University. 4- Collins, S.L. Sinsabaugh, R.L. Crenshaw, C. Green, L. Porras Alfaro, A. Stursova, M. Zeglin, L.H (2008). Pulse dynamics and microbial processes in aridland ecosystems, Journal of Ecology, 96. 5- Dingman, S. L., Seely-Reynolds, D.M., and Reynolds, R.C (1998). Application of Kriging to Estimating Mean Auual Precipitation in a Region of Orographic Influence. Journal of the American Water Resources Association, 24. 6- Donald L Phillips L,D. Jayne Dolph J and Marks, D (1992). A Comparison of Geostatistical Procedures for Spatial Analysis of Precipitation in Ountainous Terrain. Agricultural and Forest Meteorology, 58. 7- Dehghani, M (1387). Application and an estimate of the types of kriging, MA Seminar, University of Yazd. 8- Goovaerts, P (2000). Geostatistical Approach for Incorporating Elevation into Spatial Interpolation Rainfall. Journal of Hydrology, Amsterdam. 228(1-2). 9- Hargrove, W (2001). Interpolation of rainfall in Switzerland using a regularized spline with tension. Geographic Information and Spatial Technologies Group, book Ridge National laboratory. 10- Iran Meteorological Organization (1388). Meteorological Yearbook. 11- Isaaks, E.H, Srivastava, R.M (1989). Applied Geostatistics. Oxford University Press: New York. 12- Kalantari, Khalil (1385). Processing and data analysis in social research - Economic using software SPSS, Sharif publication. 13- The Assessment of Geostatistic Methods and Linear Regression in ... Knapp, A.K. Smith, M, D (2001). Variation among Biomes in Temporal Dynamics of Aboveground Primary Production, Journal of Science, 291. 14- Krishna murthy, B.R and SAbbaiah, G (2007). Geostatistical Analysis for Estimation of Mean Rainfalls in Andhra Pradesh, India, International of Journal of geology, 3(1). 15- Lookingbill, T.R and Urban, D.L (2003). Spatial estimation of air temperature differences for landscape-scale studies in montane environments. Agricultural and Forest Meteorology, 114. 16- Lashany Zand, Mehran (1381). The guidelines deal with drought and climate of Iran (case study of six basins in West and North West of Iran), Ph.D. Thesis, University of Isfahan. 17- Madani, H (1377). Fundamentals of Statistics, Amirkabir University Press. 18- Masoudian, S. A (1382). Analysis of monthly temperature structure of Iran, Isfahan University Journal. 19- Mehdizadeh, M (1381). Geostatistical estimation of temperature and precipitation in the lake basin, MA thesis, Agricultural Meteorology, College of Agriculture, Tehran University. 20- Pak Hosni, AA (1380). Exploratory data analysis, Tehran University Press. 21- Soltani, S., Modarres, R (1385). The frequency and intensity of meteorological drought in Isfahan, Iran Journal of Natural Resources. 22- Shams al-Din, A (1379). Changes in regional precipitation using kriging in the Northern provinces, irrigation and drainage MA thesis, Faculty of Agriculture, Shiraz University. 23- Safari, M (1381). Determine the optimal network of groundwater level measurements using geostatistical methods, case study Chmchal Plains, irrigation and drainage MA thesis, Faculty of Agriculture, Tarbiat Modarres University. 24- Ronald, P.B., and jay, M.V.H (2009). Blackbox Kriging: Spatial Prediction without Specifying Variogram Models, Journal of Agricultural, Biological, and Environmental Statistics, 1(2):297-322. 25- Tazeh, M., Kothari, MR; Bkhshaei, M and Khosravi, y (1387). The land zoning Transo index using the statistics and GIS (Case study: the western part of the Esfahan province), International Conference botanical trees and ecosystems to climate change in the Caspian, Caspian Institute Sari ecosystems. 26- Tally Ghahrood, Manijeh (1384). Three-dimensional GIS environment, Publications Unit, Teacher Training University Jihad, Tehran. 27- Tsakiris G, Vangelis H (2004). towards a drought watch system based on spatial SPI. Water Resour Manag 18(1). 28- Wang,F (2006). Quantitative methods and applications in GIS. CRC Press.
Keywords
