Calculation and verification of the irrigation probability index using rain-fed indicator line

Abstract: The information of cultivated land irrigation area, distribution, irrigation amount and irrigation time is of great significance to food security, economic development and water resources management. Among them, irrigation area and irrigation distribution are the most basic irrigation information, which can be obtained through irrigation farmland mapping. Most current researches determine the irrigation indirectly by the parameters reflecting vegetation growth or soil moisture, and the physical mechanism is unclear. There is a high demand to develop irrigation characteristic parameters with clear physical meaning and strong indicative significance to further improve the accuracy of irrigation farmland mapping. Based on the principle that irrigation can slow down or inhibit the evolution of meteorological drought to agricultural drought, this study proposed the concept of rain-fed indicator line and develop the irrigation probability index. Nebraska, which has a good irrigation data base, was selected as the test area. At first, it was assumed that in areas with the same meteorological drought intensity, the severity of agricultural drought was closely related to irrigation. We used precipitation, actual evapotranspiration and potential evapotranspiration to calculate Crop Water Deficit Index (CWDI) and Crop Water Stress Index (CWSI). Then, the CWDI was used to characterize the meteorological drought, whereas, the CWSI was to characterize the agricultural drought in the study area. Taking the CWDI of farmland pixel as the x-axis and CWSI as the y-axis, the two-dimensional scatter map of farmland pixel was made. The upper envelope of scattered points was extracted as the rain-fed indicator line, where the distance from each farmland pixel to the rain-fed indicator line along the vertical coordinate was defined as the Irrigation Probability Index (IPI). The relationships were then determined between the IPI and real irrigation area, the number of active irrigation wells, the area of irrigation facilities, and the distribution of rivers. The results show that the IPI of Nebraska was higher in the east and lower in the west, which is consistent with the distribution pattern of temperature, precipitation and agricultural activities in the state. There was the highest correlation coefficient (0.62) between the sum of county IPI and the real irrigation area, followed by the area of irrigation facilities (0.55), and the lowest with the number of active irrigation wells (0.51). Three correlation coefficients all passed over the test of 0.05 significance level, indicating that the IPI can effectively characterize the possibility of farmland to be irrigated. However, there was different adaptability of IPI in three climate regions (humid, climate suitable, and arid region). Specifically, the IPI can better characterize the possibility of farmland to be irrigated in the climate suitable area and the arid area, compared with the humid area. Particularly, the sum of county IPI in climate suitable and arid areas presented a significant positive correlation with the real irrigation area, the number of active irrigation wells, and the area of irrigation facilities. The average IPI decreased with the increase of the distance to the river. There was no significant correlation between the sum of county IPI and the number of active irrigation wells in the humid area, where the average IPI increased with the increase of distance to the river. The newly-developed IPI in this study can be widely expected to better represent the possibility of farmland to be irrigated, further serving as a physical parameter for the irrigation farmland mapping.