2014-05-15_REVISION - M1987064 - Laserfiche WebLink

Home Browse Search

D. C. Goodrich et al. /Agricultural and Forest Meteorology 105 (2000) 281 -309 demonstrated that specific yield was the least sensitive parameter in their localized water balance computa- tion around Lewis Springs. A similar computation for change in storage was made over DOY 101 -191. The uncertainty in this term was estimated by assuming the errors in the areas of the alluvial aquifers were up to 15 %, the change in water table measurement errors were roughly 5 mm and the specific yield error was 0.05. Soil moisture storage changes were also scaled using land cover class areas. The integrated change in soil moisture at the mesquite tower was multiplied by the high- and low - density mesquite land cover areas. The soil moisture changes at the sacaton tower were assumed to apply to the sacaton, scrub and bare land cover class areas and the changes at the mesquite tower were assumed to apply to the C/W areas. Based on calibration of the soil moisture instruments an error of 2% of volumetric soil moisture was assumed. The er- ror in the depth of probe placement was assumed to be lcm. The changes in water table depth were determined by selecting a set of piezometers within each of the aquifers and averaging the change in depth from the beginning to end of the time period examined. The piezometers selected to determine the change in head level for the pre - entrenchment aquifer were WNF 13, WMF 13, WSF 14, WMC 14, and ENC 13. For the post - entrenchment aquifer, piezometers EMC 10, EMF 10, ESC 10 and ESF 11 were used (see MacNish et al., 2000 for more detail). Using this information and Eq. (8), GWnet(80--90) (DOY 80-90) was 73 200 m3 which is equivalent to a rate of gain of 0.0847 m3 s -1 over the 10 days period. For the same period of time, the average gradient between BLM well #4 and BLM well #5 was computed (V(4,5)(80-90)). These wells are located on cross - section 3 at Lewis Springs with well #4 finished at a depth of roughly 7m and well #5 finished at a depth of roughly 54 m. The net groundwater gain during the water balance period from DOY 101 to 191 was computed by ad- justing the pre - green -up gain with the ratio of the ob- served to pre - green -up gradient between wells 4 and 5 as follows: GWnet(101 -191) 191 j 0(4 5)i 1 GWnet(80 -90) L j (10) i =1oi V(4> 5)(go -9o> 295 This was done in an attempt to account for changes in the net groundwater gain resulting from time - varying changes in the regional aquifer. The uncertainty asso- ciated with this term was computed by substituting the standard error for the Lewis Springs stage — discharge rating for GWnet(80--90) in Eq. (10). Error in difference in head measurements and the separation distance be- tween the deep and shallow well were assumed to be 0.005 and 0.5 m, respectively. The volume of water transpired by the C/W forest (Tc/wr) was estimated by summing the average daily values of transpiration obtained from the calibrated P —M model scaled by the remotely sensed area for the riparian woodland (Column 2 of Table 1). Scal- ing in this fashion implies several assumptions: (1) the meteorological measurements made at the Lewis Springs mesquite tower are representative of the entire reach; (2) the trees as well as the relative proportions of young and old trees sampled at Lewis Springs are representative of the entire reach. The last assumption was required as the remotely sensed data were not able to distinguish old from new growth. The uncer- tainty in this term is relatively large due to the multiple measurements and estimates that are made in scaling from the tree, to stand, to riparian corridor level. Er- rors in the scaling process include: the standard error in the sapwood area versus tree diameter regression, the standard error of the (sap flux/unit sapwood area) measurements, the daily error in the P —M model cali- bration, and the estimated error in the remotely sensed area of riparian woodland (2 %, see Appendix A). The volume of water evaporated and transpired by the mesquite trees (ETm) was determined by summing the average daily values of mesquite ET multiplied by the area for this vegetation class. As noted in Ap- pendix A, the uncertainty associated with the area of this vegetation class is estimated to be 2 -5% and the larger figure was used in the error calculation. The Bowen ratio ET estimates are assumed to be accurate to within 20 %. 5. Results and discussion Several preliminary results regarding sacaton grass ET water sources and the ability of remotely sensed data to detect changes in the hydrologic regime will be briefly discussed. More substantial results are