2014-05-15_REVISION - M1987064 - Laserfiche WebLink

Home Browse Search

D.C. Goodrich et al. /Agricultural and Forest Meteorology 105 (2000) 281 -309 4. Theory and approach The riparian ET experiments were designed to ob- tain representative measurements at the Lewis Springs site over the primary vegetation types and spatially scale these measurements using remotely sensed data over a larger portion of the corridor. Temporal scal- ing (or extrapolation) between synoptic measurement periods was to be obtained either from continuous ET measurements over the various vegetation types or from a meteorologically driven model. The follow- ing approaches will be presented: (1) a methodology for scaling the ET estimates in time and space; (2) a modeling approach to temporally scale (extrapolate) the C/W transpiration estimates; (3) a water balance over a 10 km reach of the riparian corridor to check the riparian ET estimates. 4.1. Scaling approach The scaling approach used was dependent both on the type of riparian vegetation and the measurement method employed. Continuous measurements of the local meteorology and several of the energy balance components were made through the growing season over relatively uniform areas of the sacaton grass and mesquite. The Bowen ratio technique was used with these data used to estimate ET [LT -1 ] in for these veg- etation types (Scott et al., 2000). The remotely sensed estimates of sacaton and mesquite area were then used to scale these measurements spatially over portions of the riparian corridor to obtain ET as a volume per unit time. Because continuous measurements were avail- able, these estimates could simply be integrated over the desired time period to obtain a volume of ET from mesquite and sacaton for the overall water balance. Scaling the C/W transpiration estimates obtained from sap flux measurements presented additional challenges. With tree coring, the basic data collected consisted of sap flux (tree transpiration), expressed on a per unit sapwood area basis per unit time (g H2O cm -2 h-1). The first step to scale individual tree estimates of transpiration to patch and stand es- timates involved relating sapwood area to an easily measured parameter. In this case, basal tree diameter was related to sapwood area by a separate regres- sion analysis for C/W trees (Schaeffer et al., 2000). In the next phase of scaling, 12 C/W forest patches 287 (five newly established and seven successionally ad- vanced) were selected based on ease of identification on the remotely sensed imagery and high - resolution CIR photography (see Fig. 1, Schaeffer et al., 2000). A number of the patches were also selected as they were within the LIDAR data acquisition area (Cooper et al., 2000). The areal canopy area of each of the patches was measured using the aerial imagery. The basal diameters of all trees in each patch were then measured, as well as the leaf area index using a light extinction meter. The next step in scaling the sap flux measurements was from the patch level to a spatially continuous C/W stand on a river reach roughly one -half km long at the Lewis Springs site. By scaling to the level of a larger river reach, a more representative sample of trees was obtained and the uncertainty in patch area definition was reduced. Canopy area definition from outlining the perimeter of smaller patches on digital imagery is made difficult by their irregular shape and effects from shadowing. To overcome this problem and pro- vide stand level transpiration estimates, breast height diameters of all C/W trees were measured in a roughly 600 m stream reach at Lewis Springs (2238 trees) and a roughly 1150m stream reach at the Escapule site (256 trees). By including larger continuous stands of C /W, the perimeter to area ratio of the canopy and pre- sumably the uncertainty in defining the canopy area was reduced. With these data, stand level C/W transpiration es- timates were obtained continuously for each synoptic measurement period (SMP) during the times the heat pulse velocity sensors were installed in the trees. These estimates were obtained by first computing an average sap flux per unit sapwood area for each tree with installed sensors. The average sap flux per unit sapwood area times the sapwood area of all 2238 trees in the 600 m reach provided the volume of stand level sap flux for 30 min averages during each synop- tic measurement period. This volume of sap flux as a function of time is divided by the remotely derived area of the C/W canopy to provide a C/W transpi- ration flux in units of [LT -1]. During the periods of sap flux measurements, this quantity can be spatially scaled further by multiplying it by the remotely esti- mated canopy area over any specified river reach. This does not solve the problem of temporal scaling or ex- trapolation of the measurements between the synoptic