Agro-hydrological modelling of regional irrigation water demand

Abstract

The irrigation sector accounts for over 70% of the total freshwater consumption in the world. Therefore, e cient management of irrigation water is essential to ensure water, food, energy and environmental securities in a sustainable manner; these securities are grand challenges of the 21st century. The main objective of this research is to evaluate the simulation of irrigation water demand at the catchment scale in order to develop improved tools for conducting quantitative planning and climate change studies. Irrigation water demand is mostly driven by soil moisture. It is a state variable which is used to trigger the irrigation in hydrological models. In this study, a hydrolgical model (Soil and Water Assessment Tool, SWAT) is evaluated for reliably simulating the spatial and temporal patterns of soil moisture at a catchment scale. The SWAT simulated soil moisture was compared with the indirect estimates of soil moisture from Landsat and Time-domain re ectometry (TDR). The results showed that the SWAT simulated soil moisture was comparable with the soil moisture estimated from Landsat and TDR. Secondly, the applicability of the SWAT model was tested for simulating stream ow, evapotranspiration (ET) and irrigation water demand for four di erent agro-climatic zones (Mediterranean, Subtropical monsoon, Humid, and Tropical). Two di erent irrigation scheduling techniques were used to simulate irrigation namely, soil water de cit and plant water demand. It was seen from the results that the SWAT simulated irrigation amounts under soil moisture irrigation scheduling technique were close to the irrigation statistics provided by the state. However, the irrigation amounts simulated under the plant water demand irrigation scheduling technique were underestimated. Additionally, the two reanalysis data were also used to check the data uncertainty in simulating irrigation water demand. SWAT model code was modi ed by incorporating modi ed root density distribution function and dynamic stress factor. The modi ed model was used to simulate irrigation and crop yield. It was tested against the irrigation and crop yield simulated by Soil Water Atmosphere Plant (SWAP) model and eld data (Hamerstorf, Lower Saxony, Germany). It was then validated for di erent catchments (Germany, India and Vietnam). The results showed that the SWAT simulated irrigation water demand in case of plant water demand is comparable with the amount simulated by the model under soil water de cit irrigation scheduling technique. This dissertation not only bridges the gap between the scales of soil moisture determination but also establishes a close connection with the actual observations and modelled soil moisture and irrigation amounts at the eld, regional and global studies in agricultural water management. Additionally, the studies about simulating irrigation water requirement in data-scarce areas must address data uncertainty when using reanalysis data. It was found that rainfall is not always the dominant variable in irrigation simulation. Therefore, it is worth checking and bias correct the other climate variables.