Home
Soil moisture mapping
Importance of soil moisture in water resources applications
Although soil moisture constitutes a minute portion of global water resources, it plays a critical role in the hydrological cycle and directly and indirectly affects climatological and biogeochemical cycles. The water content of a soil both impacts and is impacted by the
hydrological cycle of a field, catchment, or region. For example, soil moisture (together with other factors such as slope and land use) determines the infiltration rate of rainfall and, thus, determines the proportion of rainfall that will enter the soil (eventually reach ground water) or that will be lost to the system by runoff (to surface water). Knowledge of soil moisture is particularly crucial for early warning of floods. Floods are generated when either
(i) the intensity of a long-duration rainfall event exceeds the infiltration capacity of a soil (itself determined by soil moisture) or
(ii) when rain falls on an already very wet soil. In both cases,
knowledge of the initial soil moisture combined with the forecast of rainfall events and knowledge of physical properties of the soil would be useful to anticipate floods. Updated soil moisture information for use in distributed rainfall–runoff models can be useful for forecasting
large floods as the updated soil moisture can be used as an initial condition for the forecasts
Methods for obtaining soil moisture data
Routinely measured soil moisture has the potential to significantly improve our ability to model hydrologic processes. There are three general approaches to measuring soil moisture: in situ or point measurements, soil water models, and remote sensing (Schmugge et al. 1980). A multitude of direct and indirect measuring techniques are used with these three approaches that range from the use of gravimetric, nuclear, electromagnetic, tensiometric, and hygrometric techniques to remote sensing techniques (Zazueta & Xin 1994). Each method has advantages and limitations. Schmugge et al. (1980) stated that all methods should meet three requirements:
frequent observations; an estimate of moisture within the top 1–2 m of soil; and a description of moisture variations over a large expanse, such as a county or state. The most accurate of soil moisture measurements are in situ methods, which include gravimetric, nuclear, and electromagnetic techniques. However, in situ methods can be time consuming, labor intensive, expensive, and destructive to the soil profile. Moreover, in situ measurements are point soil moisture measurement techniques, which require a better understanding of scaling, aggregation, and disaggregation in both temporal and spatial domains, as they are only accurate at the point of measurement (Schmugge et al. 1980), and are often limited to small fields with similar soil characteristics and vegetative conditions (Mohanty & Skaggs 2001). GIS can be used to
interpolate these point data to create continuous surface maps of soil moisture. However, regional scale estimation of soil moisture using in situ field observations is not reliable due to problems of representative sampling and cost (Lakshmi et al. 1997), and the large spatiotemporal variability soil moisture naturally exhibits (Leese et al. 2001). Alternatively, soil moisture can be estimated using hydrological models for regional scale studies (Uddameri & Kuchanur 2007). However, this type of modeling approach is limited by the level of accuracy desired and on the spatial scale of the required estimations. The advantages of using models are that if the models are appropriate (a big if), soil moisture can be obtained inexpensively in almost real time; if rainfall and evapotranspiration forecasts are available, the models can be run in predictive mode and provide early warning for floods and droughts. However, modeling soil moisture and soil water movement is not easy, and three
types of models can be identified:
(i) empirical models,
(ii) capacitance-based models, and
(iii) physical-based models. In principle,
any hydrological model can be used for soil moisture mapping; however, simple soil water models are useful for regional scale modeling where extensive data are lacking. Complex models require extensive data to model processes affecting water dynamics, hence not suitable for regional scale applications where such data
are lacking (Ranatunga et al. 2008). In addition, simple soil water models do not resolve spatial variations in saturation, nor do they express soil and plant behaviors as functions of climate, soil, and vegetation characteristics (Guswa et al. 2002). Therefore, a very simple model may not adequately represent the process to be modeled at a regional scale. Complex soil water models require large amounts of data for modeling and calibration including rainfall, solar radiation, air temperature, air humidity, wind speed, and soil physical properties, as well as satellite-based vegetation and LULC mapping. These models are expensive to produce and are useful only for small areas because of the increased risk of error propagation and uncertainty (Hollinger & Isard 1994; Pauwels et al. 2002). It is sometimes confusing and difficult to choose the right soil water model for a specific purpose (Ranatunga et al. 2008) since complex soil water models have given mixed results in some studies (Bernier 1985; Beven 1989; Grayson et al. 1992; Wigmosta et al. 1994) and in other studies have shown that there is no significant difference in the results between a simple model versus a more complex model (Grayson & Woods 2003; Kandel et al. 2005)
Importance of soil moisture in water resources applications
Although soil moisture constitutes a minute portion of global water resources, it plays a critical role in the hydrological cycle and directly and indirectly affects climatological and biogeochemical cycles. The water content of a soil both impacts and is impacted by the
hydrological cycle of a field, catchment, or region. For example, soil moisture (together with other factors such as slope and land use) determines the infiltration rate of rainfall and, thus, determines the proportion of rainfall that will enter the soil (eventually reach ground water) or that will be lost to the system by runoff (to surface water). Knowledge of soil moisture is particularly crucial for early warning of floods. Floods are generated when either
(i) the intensity of a long-duration rainfall event exceeds the infiltration capacity of a soil (itself determined by soil moisture) or
(ii) when rain falls on an already very wet soil. In both cases,
knowledge of the initial soil moisture combined with the forecast of rainfall events and knowledge of physical properties of the soil would be useful to anticipate floods. Updated soil moisture information for use in distributed rainfall–runoff models can be useful for forecasting
large floods as the updated soil moisture can be used as an initial condition for the forecasts
Methods for obtaining soil moisture data
Routinely measured soil moisture has the potential to significantly improve our ability to model hydrologic processes. There are three general approaches to measuring soil moisture: in situ or point measurements, soil water models, and remote sensing (Schmugge et al. 1980). A multitude of direct and indirect measuring techniques are used with these three approaches that range from the use of gravimetric, nuclear, electromagnetic, tensiometric, and hygrometric techniques to remote sensing techniques (Zazueta & Xin 1994). Each method has advantages and limitations. Schmugge et al. (1980) stated that all methods should meet three requirements:
frequent observations; an estimate of moisture within the top 1–2 m of soil; and a description of moisture variations over a large expanse, such as a county or state. The most accurate of soil moisture measurements are in situ methods, which include gravimetric, nuclear, and electromagnetic techniques. However, in situ methods can be time consuming, labor intensive, expensive, and destructive to the soil profile. Moreover, in situ measurements are point soil moisture measurement techniques, which require a better understanding of scaling, aggregation, and disaggregation in both temporal and spatial domains, as they are only accurate at the point of measurement (Schmugge et al. 1980), and are often limited to small fields with similar soil characteristics and vegetative conditions (Mohanty & Skaggs 2001). GIS can be used to
interpolate these point data to create continuous surface maps of soil moisture. However, regional scale estimation of soil moisture using in situ field observations is not reliable due to problems of representative sampling and cost (Lakshmi et al. 1997), and the large spatiotemporal variability soil moisture naturally exhibits (Leese et al. 2001). Alternatively, soil moisture can be estimated using hydrological models for regional scale studies (Uddameri & Kuchanur 2007). However, this type of modeling approach is limited by the level of accuracy desired and on the spatial scale of the required estimations. The advantages of using models are that if the models are appropriate (a big if), soil moisture can be obtained inexpensively in almost real time; if rainfall and evapotranspiration forecasts are available, the models can be run in predictive mode and provide early warning for floods and droughts. However, modeling soil moisture and soil water movement is not easy, and three
types of models can be identified:
(i) empirical models,
(ii) capacitance-based models, and
(iii) physical-based models. In principle,
any hydrological model can be used for soil moisture mapping; however, simple soil water models are useful for regional scale modeling where extensive data are lacking. Complex models require extensive data to model processes affecting water dynamics, hence not suitable for regional scale applications where such data
are lacking (Ranatunga et al. 2008). In addition, simple soil water models do not resolve spatial variations in saturation, nor do they express soil and plant behaviors as functions of climate, soil, and vegetation characteristics (Guswa et al. 2002). Therefore, a very simple model may not adequately represent the process to be modeled at a regional scale. Complex soil water models require large amounts of data for modeling and calibration including rainfall, solar radiation, air temperature, air humidity, wind speed, and soil physical properties, as well as satellite-based vegetation and LULC mapping. These models are expensive to produce and are useful only for small areas because of the increased risk of error propagation and uncertainty (Hollinger & Isard 1994; Pauwels et al. 2002). It is sometimes confusing and difficult to choose the right soil water model for a specific purpose (Ranatunga et al. 2008) since complex soil water models have given mixed results in some studies (Bernier 1985; Beven 1989; Grayson et al. 1992; Wigmosta et al. 1994) and in other studies have shown that there is no significant difference in the results between a simple model versus a more complex model (Grayson & Woods 2003; Kandel et al. 2005)