Introduction:

Over the next several decades, climate change is likely to affect the distribution and activity of arctic plants and animals in ways that are largely unknown at this time. Due to the strong effect of temperature on the processes of photosynthesis and respiration, it is likely that warming will alter the carbon balance of the region. Altered climate may also affect species composition, which could lead to further changes in ecosystem processes.

This document summarizes early findings of our study in Barrow, Alaska, which began in June, 2000. We are applying novel optical sampling methods to enhance our understanding of the biological impacts of climate change in arctic ecosystems. Because climate impacts are likely to occur over large regions, it is essential to apply these optical sampling tools over large areas from a distance ("remote sensing"). Because many existing remote sensing methods suffer from technical limitations in northern latitudes, we are exploring improved sampling methods that employ the following features:

1) Novel optical sensors that allow measurement under all weather conditions.

2) Hyperspectral (narrow-band) detectors that detect changing species composition and physiology.

3) Thermal sensors that sample surface temperature, an important property affecting gas fluxes.

4) Multi-scale sampling, enabling us to link optical properties to physiological states at different spatial scales (leaves to large landscape regions).
 

Multi-Scale Sampling:

This project samples optical properties at various spatial scales, from individual organisms to large landscapes. Sampling at finer scales enables us to link optical traits to species composition and physiological activity, providing a solid basis for interpreting remotely sensed signals at progressively larger scales. Sampling at larger scales allows us to extend our understanding to large regions.
 
 

Figure 1. When viewed closely, the tundra is a miniature landscape of lichens, and vascular plants. Individual organisms can be distinguished by their optical properties. The contrasting reflectance patterns of the different "functional types" (lichens, mosses, and vascular plants) shown in this image reveals their varying pigment composition associated with contrasting physiological function.

Figure 2. Spectral reflectance reveals contrasting pigment composition and physiological activity for different cover types (lichens, vascular plants, and mosses). Shown here are spectra for representative cover types shown in figure 1.

Figure 3. Optical and physiological sampling of individual species reveal contrasting functional behavior of lichens, mosses, and vascular plants. Relative to mosses and vascular plants, lichens show a reduced photosynthetic light-use efficiency, chlorophyll fluorescence yield (?F/Fm') and greenness (NDVI). This illustrates the potential for altered species composition to change ecosystem function.

Figure 4. Treatment plots (where temperature and water levels have been manipulated) allow us to relate surface reflectance to altered ecosystem function. Shown here is a heated plot. Treatments provided by Glen Kinoshita (SDSU).

Figure 5. Reflectance of treatment plots reveals a clear effect of heating on surface greenness (NDVI), suggesting altered ecosystem carbon fluxes. Overlaid on the treatment plot data, are the transect data (see figure 8, below). (Heating treatments provided by Glen Kinoshita, SDSU.)

Figure 6. Transect sampling provides repeatable measurements of the tundra, allowing us to monitor temporally changing surface properties, and revealing the effect of microtopography, species composition, and physiological activity on surface reflectance. The movable cart in this picture supports radiation sensors that sample surface reflectance, surface temperature, and sky irradiance along a distance of 100 m.

Figure 7. Example of transect reflectance collected near the beginning (July 7) and middle (July 23) of the growing season, illustrating different greening rates for different tundra locations. Note also that the lowest regions (dominated by vascular plants) tend to be more productive (shown as having highest NDVI values on July 23). By contrast, the higher regions (having more lichen cover) tend to be less productive (low NDVI values).

Figure 8. The seasonal trend of the greenness index (NDVI) shows a steady increase in vegetation greenness from snowmelt (mid-June) to peak season (mid-August). Each point represents the average of 100 transect sampling points.

Figure 9. An initial comparison of surface greenness (NDVI sampled along transects) to ecosystem fluxes (sampled via eddy covariance) yields a strong correspondence between ecosystem carbon uptake and greenness (NDVI) early in the growing season. (CO₂ flux data courtesy of Hyojung Kwon, SDSU.)

Figure 10. Aircraft sampling of surface optical properties from the Sky Arrow mobile flux platform extends our understanding of surface properties to large landscape regions.

Figure 11. Reflectance measurements from our first flight this summer (July 2000) illustrate changing greenness (Normalized Difference Vegetation Index) and water content (Water Band Index) associated with fine-scale topography and cover of the tundra surface. Total transect distance was approximately 3 km. Further flights will link these patterns to surface photosynthetic rates over larger distances. (Aircraft data courtesy of Rommel Zulueta)

Conclusions:

1) There is a strong link between optical properties and ecosystem CO₂ fluxes, demonstrating that remote sensing can reveal changing ecosystem physiology in the arctic),

2) Warming alters ecosystem optical properties, demonstrating that remote sensing can detect biological responses to warming), and

3) Different cover types (lichens, mosses, and vascular plants) have contrasting optical properties and photosynthetic rates, indicating the potential for changing species composition to alter regional carbon balance.
 
 

Future Plans:

Further work through 2001 will attempt to derive a quantitative model that can predict photosynthetic fluxes from remote sensing of arctic ecosystems. We will test this model at the various spatial scales indicated here.
 
 

For further information:

visit http://vcsars.calstatela.edu
 
 

Acknowledgements:

This project is supported by funding from IARC, and is a collaboration with San Diego State University scientists (Dr. Walter Oechel and colleagues) who have generously contributed to this effort. Additional support of BASC staff and facilities has also been instrumental to the success of this project.