Irrigation water management in arid regions of Middle East: Assessing spatio-temporal variation of actual evapotranspiration through remote sensing techniques and meteorological data

摘要

Spatial and temporal distribution of reference (ETo) and actual evapotranspiration (AET) over the central region of Saudi Arabia during 1950-2013 are estimated using remote sensing and GIS techniques. Firstly, the FAO Penman-Monteith method was used to model the spatial distribution of ETo on a grid-by-grid basis using data collected from meteorological stations and GIS techniques. Then, crop coefficients (Kc) were modeled as a function of 16-day time-series MODIS normalized difference vegetation index (NDVI). Next, using Kc maps and ETo as input, daily AET was simulated by the soil water balance (SWB) model and aggregated to monthly and annual AET. From empirical NDVI-Kc relationships developed and applicable at pixel level, Kc derived from the NDVI-Kc relationships agree well with Kc recommended by FAO over various crop growth stages in the field. The monthly AET maps for 1950-2013 show a gradual increase in AET during the crop-growing season in January to May but a subsequent decline as the season progresses from June to December. The AET estimated for January to June are arranged in descending order, which are May (3.67-44.7 mm/day), April (5.99-36.8 mm/day), March (2.96-32.7 mm/day), February (0.68-20 mm/day), June (2.42-17.7 mm/day) and January (1-11 mm/day), respectively. Statistical analysis shows that statistically significant change point in daily AET generally occurred in 1990, such that the long-term average daily AET of 1950-1990 at 3.6 mm/day increased to about 5.3 mm/day between 1990 and 2016 with a positive trend of 1.5 mm/decade. The annual AET estimated for irrigated cropland in northern and central regions of Riyadh, Al-Qassim province and Hail province range from 1200 to 2900 mm/year. In these regions, low AET values are found in shrubland, grassland, and other natural vegetation. The annual AET estimated by the SWB model are about 9-11% higher than modeled AET in the study area, where the long-term average daily AET estimated for 1950-2013 range from 2 mm/day to 30 mm/day. Representative AET maps derived from applying the NDVI-Kc relationships to the SWB model will be useful to achieve the planning and management of sustainable water use in arid regions of Middle East.