PARENT SESSION
Oral Session #34: Forests: Models of function and structure. Presiding: M. Lechowicz.
Tuesday, August 7, 2001. 1:00 PM to 3:45 PM. Hall of Ideas F.
Short-term prediction of transpiration from managed forests in northern Wisconsin.
Mackay, David1, Ewers, Brent1, Ahl, Douglas1, Samantha, Sudeep 1, Gower, Stith1, 1
ABSTRACT- Land use changes over the last century have resulted in a mosaic of forest types in northern Wisconsin. We developed a landscape scale coupled forest ecosystem - hydrology model to examine the role of changing forest cover types on land surface - atmosphere water flux. We made continuous measurement of transpiration in four cover types, red pine, sugar maple/basswood, quaking aspen/balsam fir, and northern white-cedar/balsam fir/alder, using Granier-type and Kucera-type sapflux sensors from June through September, 2000. The four cover types represent 90 percent of the basal area and 80 percent of the ground area. The other 20 percent consists of mostly non-forested areas, including grassland, shrubland and water. Transpiration from each of the stand types and their respective 7 dominant forest species were simulated with inputs of half-hourly micrometeorological measurements. Daily predictions explained from 64 percent to 92 percent of the variance of the daily measured transpiration, with some species showing tradeoffs between maximizing variance explained and minimizing bias. Multi-criteria Monte Carlo uncertainty estimation was used to examine the sensitivity of the model across the modeled species types, and to differences in vertical position (above canopy, within-canopy, and below canopy) of micrometeorological inputs. The model exhibited wide variation in sensitivities to different physiological and structural model parameters among species and stands, as well as differences in correlation between parameters. The model was also sensitive to the position of the micrometeorological measurements.
KEY WORDS: evapotranspiration, managed forests, simulation, predictive uncertainty