Identifying critical thresholds for acute responses of plants and ecosystems to water stress (TARP)

Principal investigator: Paul J. Hanson

DOE Program for Ecosystem Research

Project goal

Characterize effects of acute spring and early summer drought on current year growth, carbohydrate reserves, and canopy senescence of overstory chestnut oak (Quercus prinus) and yellow-poplar (Liriodendron tulipifera) trees.

Ecosystem being studied

An upland oak forest on Walker Branch Watershed, Tennessee. This forest is representative of temperate deciduous forests that comprise about 40% of current forest lands in the United States. Such forests are among the most productive natural ecosystems, store substantial amounts of carbon, are a key source of clean water, and are a repository of biodiversity.

An assessment of the similarity of the research site to other locations in the United States is captured in the map available at this link.


Basal area growth

Effect of treatment on cumulative basal area growth during 3 years of imposed acute drought (Lt, yellow-poplar; Qp, chestnut oak).

Through 3 years of growing-season rainfall exclusion (-100% in 2003, 2004, and 2005), large trees did not show the hypothesized responses to severe surface-soil water deficits that were indicated by previously observed stand-level sensitivity to soil drying. No consistent significant changes in tree leaf physiology, tissue carbohydrate status, or whole-tree sapflow were observed through early 2005. During the 2005 summer, however, tree growth did slow in response to the severe drought treatments.

The initial (2003 and 2004) lack of imposed-drought effect on growth was unexpected because a naturally occuring late-season drought in 1998 produced 50% reductions in forest canopy function of the trees on the adjacent Throughfall Displacement Experiment for similar levels of surface-soil drying.

The experimental manipulations did preclude normal rewetting of the surface soils in all years, but species specific differences were evident. Deep soil (35-70 cm) water extraction under yellow-poplar was lower than anticipated.

The lack of response to severe-surface soil water deficits was hypothesized to result from hydraulically efficient deep or lateral roots extending beyond the treatment area. Lateral trenching of all treatment plots in the mid-summer of 2004 had no effect on tree water use or leaf physiology leading to the conclusion that deep roots must be supplying water from deep in the soil. After the 2005 growing season, subsuface excavations of all trees were conducted to characterize the depth, number, and hydraulic conducting capacity of deep roots for each tree species. The data indicated that mean rooting depth was indeed deeper than previously assumed from root mass data collected from periodic coring near the study site.

Soil water potential

Effect of treatment on measured soil water potential in the upper 35 cm (0-35) and next 35 cm (35-70) of surface soil.

Associated with the available ground area for these experimental trees, axial conductance of deep fine-roots below 70 cm and even below 130 cm is adequate to supply the daily water demands for these trees. Radial conductance limitations will, however, place a limit on this maximum capacity and is still being evaluated. Post-treatment evaluations of the presence of mycorrhizae with soil depth are also being conducted to further characterize potential root water extraction capacities of these trees.

The study will conclude with a model-based analysis of the ability of deep root water conductance to supply and sustain large eastern deciduous trees during periods of severe soil water deficits. These results will be applicable to other large established trees, but do not characterize alternate observations of seedling sensitivity during establishment and early growth, or seedlings and saplings growth responses in the canopy understory (see also Throughfall Displacement Experiment).

Further information is available at the project's website.

Why this is important

Vertical root profiles

Biogeography models and dynamic ecosystem simulation models used in the assessment of climatic change effects on terretrsial ecosystems predict species displacement and population migration following mortality. Although such predictions are based on reasonable hypotheses regarding plant response to warming and inter-specific competition, predicted mortality rates driving current models remain largely untested. This study provides critical field data on the effect of acute drought on mechanisms responsible for growth and mortality of large deciduous forest canopy trees. The results are therefore appropriate to the evaluation of predictions of tree responses to drought.


The study uses tents to remove 100% of the growing-season throughfall and stem flow from canopy trees. Each tented, treatment area was designed to cover an area equal to or exceeding each trees projected canopy distibution at the ground surface. The tents cover an area encompassing a circular radius of approximately 10 m from the bole. Soil water content to a depth of 35 or 70 cm is meaured periodically at four positions around the control and treatment trees, and automated hourly observations are logged from a depth profile of sensors installed in one soil pit 4 m from each tree. Meausurements of soil and air temperature and surface relative humidity (1.5 m) are also logged.

Key measured response variables include: weekly dendrometer band growth observations; continuous measurement of sapflow as an integrated measure of whole-tree physiological response; periodic and campaign-based assessments of leaf photosynthesis, conductance, and osmotic potential; and nonstructural carbohydrate concentrations for terminal branch, sapwood, and fine root tissues.

The experimental system included the removel of nearest neighbor trees and saplings for both the ambient and drought treatments, resulting in a reduction in the 'stand' basal area. This reduction did allow the target trees access to more soil water storage than is present at the nearby Throughfall Displacement Experiment, and can explain (in part) why similiar surface soil drying produced alternate responses in the two studies.

TARP infrastructure

TARP tent infrastrucure around one of the treatment trees.

Key publication

Hanson PJ, Tschaplinski TJ, Wullschleger SD, Todd DE Jr., Auge RM (in press) The resilience of upland-oak forest canopy trees to chronic and acute precipitation manipulations. Proceedings of the 15th Central Hardwood Forest Conference, USDA Forest Service, General Technical Report SRS-XX. Southern Research Station, Asheville, NC, xxx-xxx

Other program publications are listed on the Publications Page.


Paul J. Hanson, Oak Ridge National Laboratory

Stan D. Wullschleger, Oak Ridge National Laboratory

Timothy J. Tschaplinski, Oak Ridge National Laboratory

Robert M. Auge, University of Tennessee, Knoxville

Funding period: February 2002 to September 2006