A water balance model for characterization of length of growing period and water stress development for rainfed lowland rice

Abstract

There is large year-to-year variation in rice production across the Mekong region (Laos, Cambodia and Thailand) due to uncertainty in the timing of the onset of the wet season and drought stress that may develop at any time during the growth of rainfed lowland rice. Unique to the nature of lowland water balance is a large component of deep percolation water loss, which depends on soil texture. The objectives of this study were to develop a soil water balance model for calculating the amount of water held in field storage (i.e. in soil and, if there is standing water, above the soil surface) and to apply it to determine the length of growing period (LGP) and water stress development in relation to soil type and rainfall pattern for the rice ecosystem. The water balance is computed separately for above-ground plus topsoil layer and subsoil layer. Components of the water balance are the existing amount of stored water, rainfall, evapotranspiration, deep percolation, and runoff. The deep percolation rate was determined from clay content in each soil layer. The model runs with daily or weekly weather data to estimate the soil water level for the growing period in the wet season. The model was validated with data collected from top, middle and bottom of rainfed lowland fields in Savannakhet province, Laos. The best correlation between the observed and simulated water level was obtained ( r 2 = 0.41) for middle fields. The simulation results showed that LGP varied greatly from year to year, particularly in locations with sandy soils, due mostly to variation in monthly rainfall occurring at the early part of the growing season (April), but also to some extent by variation at the end of growing season (October). Soil texture on the other hand is shown to have a large influence on the end of the rice growing period and hence LGP, and also water stress development during growth. Sandy soils with clay content less than 7% that are prevalent in the province are shown to cause frequent water stress and early finish in rainfed lowland rice. The model accordingly provides reasonable outputs that can provide a geographical dimension of soil hydrological patterns for various rice growing environments, and also identify the spatial pattern of drought stress that is likely to occur. Model outputs can be used to provide guidelines for practical advice to the rice farmers and researchers for determination of appropriate crop management strategies (e.g. time of planting, varieties), and policy makers for investment decisions on inputs (e.g. fertilizer price) aimed at increasing rice productivity in this Mekong region.