Eleven USAWs, ranging in size from 0.5 to 5.7 ha (1.2 to 14.1 ac), were instrumented within the LWREW to monitor smaller streams that flow into the main channel.
Use of the curve number method in turn facilitated the comparison of simulated runoff responses among the USAWs and subwatersheds, given the various land cover and soil features represented within the study area.
Table 1 lists the respective number of subbasins and HRUs that were delineated for each of the USAWs and subwatersheds within the LWREW Delineated HRUs were assumed to be spatially uniform in terms of soils, land use, topography, and climate data.
A noticeable difference between the USAWs and the subwatersheds delineated in this study was the size of the respective subbasins.
The following procedure was used to calibrate model parameters in SWAT that govern only the surface runoff response on the USAWs. The default value of the SOL_AWC was assumed to be valid for soil conditions on each of the USAWs.
A similar procedure described above for the calibration of the USAWs was used to calibrate subwatersheds 483, 442, and 526.
A comparison of measured versus simulated average runoff, PBIAS, and NSE for each of the four LWREW USAWs and the three subwatersheds is presented in table 4.
To substantiate assumed values of CN2 and SOL_AWC used for the USAWs, model simulations were performed on subwatershed 483 (4.3 [km.sup.2] [1.7 [mi.sup.2]]) and subwatershed 442 (33.3 [km.sup.2] [12.9 [mi.sup.2]]) of the LWREW.
A simulated winter wheat to pasture/range total runoff ratio of about 1.2 compares favorably to a measured ratio of about 1.3 for similar soil and land cover conditions in a long term study conducted on USAWs adjacent to the LWREW (Water Quality and Watershed Research Laboratory 1983).
As described earlier, the calibration procedure used for estimating the surface runoff response on the USAWs consisted of adjusting values of ESCO with the assumption that default values of SOL_AWC were valid and published values of CN2 were applicable for known field conditions on the LWREW.
This is because differences that exist among calibrated values of ESCO from the four USAWs reflect a degree of uncertainty that makes extension of this parameter difficult on larger watersheds such as LWREW 526.
The purpose of this investigation was to determine whether or not model parameters that govern the surface runoff response in SWAT that were calibrated from USAWs could be scaled up to provide accurate runoff simulations at a watershed scale.