Use of MIKE SHE for estimation of evapotranspiration in the Sprague River Basin

Graduation date: 2008MIKE SHE is a fully distributed, physically-based hydrologic model that can simulate water\ud movement over and under the Earth's surface. Evapotranspiration (ET) is one of the\ud components of this model. MIKE SHE uses a modification of the Kristensen -Jensen (1975)\ud method to calculate actual ET. This method is based on addition of the three\ud evapotranspiration components – interception storage, transpiration by the plant and\ud evaporation from the soil surface, to compute total actual evapotranspiration. The validity of\ud the Kristensen-Jensen method has been tested on an arid region within the Sprague River\ud subbasin of the Upper Klamath basin in southern Oregon. The model was setup on a 1,000 m\ud by 1,000 m flat surface as a one-dimensional grid cell. There are sixteen computation layers\ud which make three soil profile layers with varying soil properties. Meteorological data from\ud the Pacific Northwest Cooperative Agricultural Weather Network (AgriMet) were used to\ud setup the model. Soil physical properties were taken from the Soil Survey Geographic\ud (SSURGO) database of the Natural Resources Conservation Service (NRCS). Values of the\ud van Genuchten parameters for soil water retention and hydraulic conductivity as a function of\ud soil texture from Carsel and Parrish (1988) were applied.\ud Wetland vegetation such as duckweed and cattail, natural vegetation such as big sagebrush,\ud ponderosa pine and juniper, and agricultural crops such as grass pasture and maize were used\ud to test MIKE SHE evapotranspiration simulation. The length of growth stage, crop\ud coefficient, leaf area index (LAI) and root depth values were taken from the literature. Actual\ud crop ET rates were calculated based on AgriMet reference ET which uses the Kimberly\ud Penman (Wright, 1982) method. The alfalfa reference ET was converted to a grass reference\ud by multiplying by a factor of 0.833 (Jensen et al., 1990). The single crop coefficient method\ud was used and soil stress was accounted for using the FAO 56 method (Allen et al, 1998).\ud Simulated irrigation was applied to maize and grass to keep the root zone soil moisture close\ud to field capacity. Crop ET rates from the MIKE SHE simulation were then compared to the\ud AgriMet based ET rates, resulting in a comparison of Kristensen-Jensen method against the\ud Kimberly Penman method. Both the Kristensen-Jensen and AgriMet simulation scenarios\ud were driven by the same reference ET and the same FAO 56 basal crop coefficient.\ud Differences are therefore a function of different methods for dealing with soil moisture stress.\ud Results indicate that the MIKE SHE simulated evapotranspiration corresponds to the\ud Kimberly Penman method for the duckweed and cattail wetlands species with resulting Nash\ud and Sutcliffe (NS) efficiencies of 0.97 and 1.00, respectively. The big sagebrush, juniper, and\ud ponderosa pine species required a soil stress correction factor for the crop coefficients and the\ud results yielded NS efficiency values of 0.14, 0.59 and 0.68, respectively. Irrigation was\ud automatically turned on for maize at a 20 percent soil moisture deficit to minimize the effects\ud of water stress and the resulting NS efficiency was 0.85. For pasture, an irrigation based on\ud average monthly water deficit for pasture in Klamath was used (Cuenca et al.,1992). This\ud resulted in a NS efficiency of 0.77.\ud Each crop requires unique treatment within the model. Required vegetation parameters such\ud as crop coefficient and LAI, climatic factors such as reference ET, and soil hydraulic\ud properties need to be based on local conditions to the extent possible. It should be noted that\ud the MIKE SHE simulations were run in a one-dimensional mode which precluded accounting\ud for spatial variability or lateral flow of surface or groundwater. The simulation results indicate\ud that converting the study area into a well irrigated pasture would require application of\ud substantial amounts of irrigation water by sprinkler or flooding. Wetlands would require even\ud more water to flood the land, but would be well suited for development of regional habitat.\ud Big sagebrush, juniper and ponderosa pine survive under natural conditions but experience\ud considerable plant stress brought on by soil water deficits which limit plant production below\ud the maximum possible growth