Effects of hardwood removal on transpiration in a longleaf pine ecosystem

Background/Question/Methods

The longleaf pine (Pinus palustris) ecosystem once covered ~36 million ha of the southeastern U.S. Because this ecosystem is characterized by relatively low evapotranspiration (ET), longleaf restoration may be a useful tool for reducing forest water use and thus improving water yield in areas of the southeast susceptible to water scarcity. However, fire suppression in many remaining longleaf stands has resulted in extensive hardwood encroachment, primarily by Quercus species. Our objective was to quantify the potential of targeted hardwood removals to reduce stand-level transpiration (E_t) in remnant longleaf pine stands. The study was conducted at Ichauway, an 11,400 ha private preserve in southwestern Georgia, USA. Restoration of longleaf pine at Ichauway has focused on hardwood removal coupled with prescribed fire to control hardwood growth. To model potential effects of restoration on E_t, we measured sap-flux in seven of the most common tree species on site (two Pinus, five Quercus). We scaled tree-level sap-flux to stand-level E_tusing forest monitoring data from 860 circular plots (0.10 ha each) surveyed three times between 2001 and 2013. We focused on a subset of 32 plots located in upland habitats that have been specifically targeted for hardwood removal treatments during that period.

Results/Conclusions

We detected large specific differences in whole-tree sap-flux with the general trend as follows: P. elliottii > P. palustris ≈ mesic oaks >> upland/xeric oaks. Tree composition changed substantially in the study plots from 2001 to 2013. Overall basal area was reduced 14%. Mesic oak (e.g. Q. virginiana and Q. nigra) basal area was reduced by 55% while upland/xeric oak (e.g. Q. falcata, Q. laevis, Q. incana) basal area was reduced by 41%. Longleaf pine basal area increased 9%. Total scaled E_t declined 8% over the same period. Most of the change in E_t was due to the removal of mesic oaks. The removal of ring-porous upland and xeric oaks had little effect on E_t because these species accounted for <2% of E_t before hardwood removal. Additional work will model the effects of changes in stand structure on interception losses and total ET. Future research needs to determine whether these potential changes in E_t are realized in full, or whether compensatory behavior of remaining and regenerating trees to increased soil moisture moderates or negates the effects of hardwood removal on water budgets.