Based on the monthly meteorological data at 58 meteorological stations from 1971 to 2000 in the Lon- gitudinal Range-Gorge Region(LRGR)and the surrounding areas,six factors including yearly ET_0, maximum monthly ET_0(May),minimum monthly ET_0(December),paddy irrigation quota and total ag- ricultural irrigation quota(yearly and main irrigation period from April to June)were selected to exam- ine their spatial variability using a geostatistical approach.The results indicated that the coefficients of Moran’s I,describing the Intensity of spatial autocorrelation,were greater in longitudinal direction than in other directions;the spatial variabilities of the six parameters were mainly caused by structural factor accounting for 60.2%—87.9%;the largest variabilities of ET_0(yearly,in May and December)ap- peared in the northwest-southeast and northeast-southwest directions.Due to the summer monsoon atmospheric circumfluence from the Indian Ocean and the Pacific Ocean,and continental warm-dry current in winter,the minimum fractal dimension and the maximum spatial variability of ET_0(yearly,in May and December)occurred in southwest-northeast and southeast-northwest directions.The highest variabilities of paddy and total agricultural irrigation requirements occurred in the south-north direction due to the vapor and energy diffusion caused by corridor function of the longitudinal rivers.The minimum spatial autocorrelation was in east-west direction,which proves that the six parameters of agricultural irrigation requirement were influenced by the“corridor-barrier”function in the LRGR and its surrounding areas. Based on the monthly meteorological data at 58 meteorological stations from 1971 to 2000 in the Lon- gitudinal Range-Gorge Region (LRGR) and the surrounding areas, six factors including yearly ET_0, maximum monthly ET_0 (May), minimum monthly ET_0 (December) , paddy irrigation quota and total ag- ricultural irrigation quota (yearly and main irrigation period from April to June) were selected to exam- ine their spatial variability using a geostatistical approach. The results indicated that the coefficients of Moran’s I, describing the Intensity of Spatial autocorrelation, were greater in longitudinal direction than in other directions; the spatial variabilities of the six parameters were mainly caused by structural factor accounting for 60.2% -87.9%; the largest variabilities of ET_0 (yearly, in May and December) ap- peared in the northwest-southeast and northeast-southwest directions. Due to the summer monsoon atmospheric circumfluence from the Indian Ocean and the Pacific Ocean, and continental warm-dry current in winter, the minimum fractal dimension and the maximum spatial variability of ET_0 (yearly, in May and December) occurred in the southwest-northeast and southeast-northwest directions.The highest variabilities of paddy and total-ir. direction due to the vapor and energy diffusion caused by corridor function of the longitudinal rivers. the minimum spatial autocorrelation was in east-west direction, which proves that the six parameters of agricultural irrigation requirement were influenced by the “corridor-barrier” function in the LRGR and its surrounding areas.