Survey on Application of Fuzzy Set Theory in Estimation of Spatial Distribution of Monthly and Annual Rainfall
2004
Rahimi Bondarabadi, Sima | Saghafian, Bahram
One of important inputs of a watershed is rainfall. Interpolation of rainfall is necessary due to large scatter gauged and to some area is not available. Geostatistical methods are important owing to considering data position and spatial correlation of data. They have limitation in use, if data be not enough. Forthat, some equation (such as regression) can produce to increase number of data by study on climatology and other properties of study area. Although generated data have a little of error that it produce unequivalency in data. Using fuzzy set theory, every data (observed, generated value and obtained from an expert or imprecise) valuated by membership function. Fuzzy variogram analysis and fuzzy kriging interpolation is executed for monthly and annual rainfall in east and southeast of Iran. Cross validation technique is used for comparison and evaluation of the methods. For this in this study analysed 3 cases. The first case all of data were fuzzified. In second case, generated some of points and fuzzified these points. In the next step, increase generating data points. Variogram analysis demonstrated spatial correlation for monthly and annual rainfall in the study area. Also, fuzzy variogram showed spatial correlation with larger range. The results show that, fuzzy kriging method was the most accurate method in estimating monthly and annual rainfall. Also, increasing number of generating point, increase estimating error.
Показать больше [+] Меньше [-]