Final Report: Foliar Chemistry as an Indicator of Forest Ecosystem Status, Primary Production and Stream Water ChemistryEPA Grant Number: R825865
Title: Foliar Chemistry as an Indicator of Forest Ecosystem Status, Primary Production and Stream Water Chemistry
Investigators: Aber, John , Bailey, Scott , Hallett, Richard , Martin, Mary , Ollinger, Scott , Smith, Marie-Louise
Institution: University of New Hampshire - Main Campus
EPA Project Officer: Hahn, Intaek
Project Period: June 1, 1998 through May 31, 2001
Project Amount: $850,000
RFA: Ecosystem Indicators (1997) RFA Text | Recipients Lists
Research Category: Ecological Indicators/Assessment/Restoration , Ecosystems
The objective of this research project was to investigate the relationships between stream chemistry and canopy characteristics determined through high spectral resolution remote sensing data.
Monitoring the biogeochemical status of forest and stream ecosystems is a key component of assessing environmental quality in the northeastern United States. Scaling field studies to the regional level, and the development of spatially-continuous estimates of ecosystem parameters, requires the use of remote sensing data.
Forest canopies are the only portion of the system accessible to optical reflectance
remote sensing instruments, and, therefore, offer the most likely target surface
for monitoring forest health in this spatial mode.
Figure 1. Canopy-soil-stand interactions at the Bartlett Experimental Forest showing a) foliar nitrogen concentration in relation to forest floor C:N ratio, b) foliar nitrogen concentration in relation to aboveground net primary production, c) forest floor C:N ratio in relation to aboveground net primary production, and d) forest floor C:N ration in relation to annual net nitrification in soils.
Ongoing work by this group has shown that forest productivity, soil chemistry, and foliar chemistry at the whole stand level are tightly linked to the biogeochemical status of the forest (see Figure 1). Based on these findings (described in more detail below), this work addresses the hypothesis that stream water chemistry, averaged over time, can be predicted from foliar chemistry.
The work funded by this grant is part of a larger effort, known collectively as the White Mountain National Forest (WMNF) Mapping Analysis of Productivity and Biogeochemical Cycling (MAPBGC) Project. The WMNF, shown in Figure 2, covers 300,000 ha in northern New Hampshire, and represents much of the forested land in the northeastern U.S. Extensive field measurements throughout the WMNF have shown that strong relationships exist between whole stand level foliar canopy N, NPP and forest floor C:N ratios, and soil nitrification rates (see Figure 1). Foliar chemistry estimates for the entire WMNF were developed under funding from a NASA/TECO proposal (1996-1999). This work was done with the Airborne Visible Infrared Imaging Spectrometer (AVIRIS) flown on-board a NASA ER-2. This instrument measures at-sensor radiance from .4-2.5mm at a 10nm resolution with a spatial resolution of 17m from the ER-2 platform. Figure 3 shows an AVIRIS-based estimation of foliar nitrogen throughout the WMNF.
Figure 2. The White Mountain National Forest in Northern New Hampshire.
Figure 3. AVIRIS predicted foliar Nitrogen coverage for the WMNF.
In addition, we have established that foliar Ca can be mapped across the WMNF using remote sensing technology (see Figure 4). The relationship between reflectance data and plot-level Ca is not as strong as that for foliar N. This initial map was developed using the original plots included in the MAPBGC project.
Figure 4. AVIRIS predicted foliar Ca coverage for the WMNF.
Figure 5. Sampling scheme for Hubbard Brook, one of the intensively measured drainage basins. Yellow lines are watershed boundaries, blue triangles are stream sampling points, and red dots are plots sampled for foliar chemistry.
Plot-level foliar Ca from the sampling in this current project covered a range of values much higher than those included in the original AVIRIS foliar Ca calibration. This data has just been analyzed in the lab, and will be incorporated into a revision of the foliar Ca calibration for the entire WMNF. This present study was designed to sample stream water at a total of 46 watersheds within 5 drainage basins in the WMNF. Sampling occurred quarterly from November 1999 through August 2001. The streams were selected within five larger drainage basins to cover a range in estimated mineralogical richness. A nested approach was used to allow us to test the optimum scale at which stream water chemistry may be predicted. Error! Reference source not found. shows one of the sampled drainage basins (Hubbard Brook Experimental Forest, 10 streams). Twenty-eight plots were established at three of the five drainage basis to sample foliar nitrogen. This data was collected to verify and/or improve the performance of the AVIRIS-based estimates of foliar N. Figure 6 shows the measured plot-level nitrogen concentration at these 90 plots versus AVIRIS predicted foliar N.
Figure 6. Independent validation for predicted foliar N values.
Error! Reference source not found. shows AVIRIS predicted canopy N versus measured streamwater nitrate for Jeffers Brook and Jefferson Notch Drainage Basins. Of the five drainage basins sampled, only Jeffers Brook and Jefferson Notch were unaffected by the ice storm of January 1988. AVIRIS N is an average of all hardwood pixels within each watershed. Streamwater N is the yearly average for each watershed. There was no significant relationship between foliar N concentration averaged for both conifer and hardwood pixels and stream water nitrate.
We explored the relationship between individual species foliar nutrient status within each watershed by using individual foliar samples from field plots. Error! Reference source not found. shows that knowing individual species foliar chemistry may aid in our ability to predict stream water nitrate concentrations.
Stream water calcium concentration was predicted using the average foliar Ca concentration for beech, spruce, and sugar maple within each watershed (see Error! Reference source not found.). This indicates that foliar Ca of these species is an indicator for stream water chemistry. This predictive relationship once again emphasizes the importance of knowing individual species. This is further illustrated by Error! Reference source not found., which shows stream water Ca versus foliar Ca averaged for all species at the plot level.
An AVIRIS derived species classification has been developed for the Bartlett Experimental Forest. Fifty-six of the 224 AVIRIS bands are used in this classification (see Error! Reference source not found.). Classes are at the individual species level. This work currently is being extended to the entire WMNF. The development of a detailed species classification such as this will allow us to evaluate remotely sensed foliar chemistry on a species by species basis, as we have done here with our field data.
Nitrogen saturation in forests is a condition where cumulative nitrogen deposition reaches a threshold that exceeds the amount of N required for plant and microbial growth. This excess N leads to increased nitrification, a process that acidifies soil and leads to losses of nutrient cations (Ca2+, Mg2+, K+) from the soil. The nutrients lost from the soil leach into groundwater, streams, and lakes. These nutrient losses can create an imbalance in forested ecosystems resulting in decreased health and productivity. In addition, the excess nitrate in stream waters can have a negative impact on aquatic ecosystems and drinking water quality. It is hypothesized that large areas of forested land in the northeastern U.S. are on the verge of becoming N saturated. The ability to quickly identify which watersheds are reaching N saturation could help us to better monitor the health of our terrestrial and aquatic ecosystems. This work provides the first evidence that remote sensing technology can be employed to estimate amounts of nitrate and calcium in stream water, thereby laying the groundwork for comprehensive change detection and monitoring methodology for forested areas that are at risk of becoming nitrogen saturated.
Figure 7. Foliar N concentration versus stream water nitrate.
Figure 8. Streamwater nitrate vs. percent foliar N for paper birch, yellow birch, and red spruce on the Jeffers Brook watersheds. Foliage samples were collected within 3 of the 5 sampling areas. Jeffers Brook is the only site of the 3 that was not heavily impacted by the ice storm in January 1988.
Figure 9. Predicted vs. Observed stream water Ca. Stream Ca was predicted with the following equation: Stream Ca = .0002069(Beech Ca) + .0002482(Spruce Ca) + .0002965(Sugar Maple Ca).
Figure 10. Stream Ca vs. plot level foliar calcium averaged for all species.
Figure 11. AVIRIS derived species map.
Journal Articles on this Report : 3 Displayed | Download in RIS Format
|Other project views:||All 16 publications||3 publications in selected types||All 3 journal articles|
||Ollinger SV, Smith ML, Martin ME, Hallett RA, Goodale CL, Aber JD. Regional variation in foliar chemistry and N cycling among forests of diverse history and composition. Ecology 2002;83(2):339-355.||
||Smith M-L, Martin ME. A plot-based method for rapid estimation of forest canopy chemistry. Canadian Journal of Forest Research 2001;31(3):549-555.||
||Smith M-L, Ollinger SV, Martin ME, Aber JD, Hallett RA, Goodale CL. Direct estimation of aboveground forest productivity through hyperspectral remote sensing of canopy nitrogen. Ecological Applications 2002;12(5):1286-1302.||