Review of recent developments in soil water simulation models

Soil water is of great importance to agricultural and hydrological systems, affecting crop yields, agricultural management practices and a wide range of physical and chemical processes in soils. Many models have been developed over the years to simulate soil water, ranging from simple water balance procedures to complex deterministic models. In this paper, some of the basic concepts, strengths, weaknesses and input requirements of soil water models are reviewed. Simple budget models which require only available water-holding capacity, based on the concepts of field capacity and wilting point, and monthly average values of precipitation and potential evapotranspiration (PET) are often adequate for climatic characterizations. However, detailed deterministic models are useful for investigating more complex processes such as solute movement. soil degradation and crop growth. These models are based on Richards' equation, with an added term for root water extraction. They generally require daily or hourly meteorological data and detailed hydraulic conductivity and water potential functions for various soil layers. Meteorological data for input into models are widely available in Canada from a network of climate stations. However, the network density is sparse for those models which use the most accurate procedures for estimating PET, requiring parameters such as wind, solar radiation and/or humidity. Moreover, pedotransfer functions generally need to be used to estimate, from existing soil databases, the required soil inputs for sophisticated models. To demonstrate a practical application, a soil water model was used in conjunction with simple crop yield/water use relationships to characterize soil water regimes and crop production risks. The probability of obtaining break-even yields (P-BEY) for continuous spring wheat was calculated and mapped for the Canadian prairie region. P-BEY ranged from less than 20% in the most arid region of the prairies to between 60 and 100% in most humid areas. An estimate of P-BEY based on modelled yields was in close agreement with that based on long-term measured yields in the sub-arid region. With increasing availability of integrated multiprocess based models, this technique has potential for studying the effects of soil characteristics and crop management techniques on P-BEY.