LC-14 Abstract

The Effects of Partial Throughfall Exclusion on Forest Flammability, Productivity, Metabolism and Nutrient Cycling in an Amazon Forest

Paulo Roberto de Souza Moutinho, IPAM - Instituto de Pesquisa Ambiental da Amazonia (SA-PI)
Daniel Curtis Nepstad, WHRC (US-PI)

Drought is changing moist tropical forests in Amazonia and elsewhere. Drought severity has increased in these regions in association with El Niño events and may increase further through global warming and land use change. The consequences of this trend include potentially large effects on greenhouse gas emissions, tree mortality, reproductive biology, biogeochemistry, and forest flammability. During the 1998 El Niño episode, for example, approximately one third of Amazon forests experienced depleted soil moisture and may have become susceptible to fire. An integrated understanding of the responses of moist tropical forests to severe drought has eluded the scientific community in part because intensive field studies of episodic natural drought “experiments” are extremely difficult.

In this research, we continue a large-scale throughfall reduction experiment in an Amazon forest. Predicted responses to a total of 5 years of experimentally-induced drought and 2 years of release from drought stress will be tested.  The study exploits the research infrastructure previously established with NSF funding, including one-hectare control and treatment study plots with canopy towers (8), catwalks (200 m), 12-m soil shafts (10) and perimeter trenches (800 m). The plots were compared during a one-year calibration period prior to treatment initiation.  Measurements include stem and root sapflow, photosynthesis, leaf water potential, xylem cavitation, fine root dynamics, soil water, canopy phenology, stem growth, stem respiration, nutrient fluxes through the forest canopy, floor and soil, and soil fluxes of CO2, CH4, N2O and NO. Beginning in 2000, throughfall has been partially excluded from the treatment plot with 5,660 plastic panels that are installed in the understory during the rainy season and removed during the dry season.

Forests growing on deep, clay soils throughout much of Amazonia are buffered against the effects of moderate droughts by the large amount of water stored in the rooting profile.  But even mild drought elicits important forest responses.  After excluding 890 mm of throughfall in 2000 and 680 mm in 2001 from the one-hectare treatment plot, several unanticipated forest responses were documented that would have been very difficult to detect without a throughfall manipulation of this magnitude. 

A reduction in soil methane emissions relative to the control plot appeared within days of installation of the exclusion panels. Surprisingly, a reduction in the growth of small tree stems (<20 cm diameter) and a reduction in sap flow were the next major treatment responses, appearing during the dry season of the first treatment year.  This decline in stemwood production of nearly 1 Mg ha-1 yr-1 is sufficient to offset the net carbon sink that has been reported for Amazon forests.  The anticipated reduction of leaf area index (LAI) appeared only during a second phase of drought in which deep soil moisture depletion began, accompanied by declines in leaf water potential. Most of the LAI reduction was caused by decreases in new leaf production instead of through increased leaf shedding. Hence the stomatal closure provoked by mild drought stress was manifested as slower stemwood growth; leaf production declined only as deep soil moisture supplies were depleted. Estimates of NPP allocation to belowground production are awaiting further analyses, but soil respiration measurements indicate little total change.

During the next three years of throughfall reduction, a third phase of drought responses is predicted in which further soil moisture depletion provokes xylem embolism and the death of trees and tree branches. Tree- and branchfall gaps will open the forest understory to sunlight, but soil moisture deficits will limit the recovery processes that gap formation usually triggers. Soil trace gas emissions will respond both to changes in soil aeration and to changes in substrate availability caused by root mortality. Movement of solutions through the soil profile will depend upon root distributions, water content, and hydraulic conductivity.  Forest floor accumulation combined with drought-induced reduction in leaf area index will greatly increase forest flammability. After cessation of the throughfall reduction treatment, persistent xylem damage in surviving large trees may restrict the supply of water to canopy foliage and suppress photosynthesis for years, while gaps are quickly filled by saplings and vegetative sprouts.

            These complex forest responses to drought, and the cessation of drought, will be tested and modeled. Field measurements will be used to test and refine canopy metabolism and forest models, improving our ability to simulate the responses of moist tropical forests to severe drought, and the cessation of drought. This experiment is the focus of 6 PhD dissertations, 2 MS theses, 5 undergraduate theses, and numerous field ecology courses, and has been widely disseminated in the popular media.