Home


       

Understanding Climate Change Effects on Glacier National Park's Natural Resources

Determining past trends and the present status of biological resources is essential for effective decision making at national parks and other lands held in the public trust. Today, however, managers have a more powerful tool for making decisions--the ability to reliably forecast resource conditions under various future scenarios. This is particularly evident when considering potential effects of climatic change on national biological resources.

   

Although predicting the future has always been an uncertain business, with support from the U.S. Geological Survey Global Change Research Program, scientists have further developed the capability for simulating the function and structure of northern Rocky Mountain ecosystems (Fig. 1). The Regional Hydro-Ecological Simulation System (RHESSys) provides quantitative estimates of key ecosystem processes for any specified point in time or space (White and Running 1994).



Fig. 1. RHESSys, the Regional Hydro-Ecological Simulation System, integrates data from various standard sources into a series of interacting models that provide quantitative estimates of ecosystem functions (modified from White and Running 1994). © Journal of Vegetation Science

A concurrently developed model, FIRE-BGC (biogeochemistry), can translate RHESSys estimates into the probable age, size, and species composition of a forest 100 years in the future on any specified mountain slope in Glacier National Park (Keane et al. 1996; Fig. 2). The predictions are spatially explicit, meaning that future forests can be calculated and mapped for each slope and aspect of the mountain landscape. The models are also mechanistic, meaning that components such as trees are "grown" in the virtual reality of computer memory by using biophysical principles and calculations, rather than by being estimated through empirical means. Finally, future landscapes can be generated by using various projections of future climates. These climatic scenarios can be imported to RHESSys from larger-scale climate modeling efforts.

   

RHESSys does not predict future climates; instead, it translates projected climate scenarios into tangible ecological changes on landscapes at Glacier National Park. Because RHESSys is spatially explicit in describing future conditions and can be displayed on a computer monitor or map, it can convey a powerfully intuitive understanding of potential landscape changes.

Fig. 2. Glacier National Park encompasses a 4,078 square-kilometer forested, mountainous landscape with numerous alpine lakes and streams.
Courtesy Glacier National Park archives
How Modeling Works  

RHESSys uses remotely sensed imagery and other satellite data to provide the geographic distribution of vegetation cover types for Glacier National Park. By combining this information with elevation data in a geographic information system (GIS), the computer can create a three-dimensional digital landscape. Estimates of biomass or rates of photosynthetic activity for each vegetation cover type in the park are made from satellite data by using other tested techniques (Running et al. 1989). RHESSys then "knows" approximately what is on the landscape and where it is; next it needs to estimate landscape response to environmental changes.

   

A microclimate model (MTCLIM) takes daily meteorological measurements and, by using mathematical expressions of physical principles, extrapolates those data to every point in the mountainous terrain (Hungerford et al. 1989). Thus, the daily changes in microclimate experienced by each stand of trees in the park is calculated. A forest biogeochemistry model (FOREST-BGC) uses the microclimate calculations (such as relative humidity or solar radiation) and appropriate biophysical principles to estimate daily tree response (Running and Gower 1991). The net result is that RHESSys can simulate forest ecosystem processes daily for many years and across large areas of Glacier National Park.

   

Of course, tree growth is determined by more than just daily weather, which is why the various models within RHESSys interact. For instance, the responses of a forest stand to changes in microclimate are passed along to calculate the effects of increasing tree growth on soil moisture. Reductions in soil moisture provide feedback to another model, which estimates streamflow in each forested drainage. Changes in soil moisture also provide feedback to the model, which estimates rates of tree growth and so on. An advantage to the RHESSys structure is that the individual models can be tested and improved independently without changing the entire simulation system. This allows RHESSys to quickly take advantage of continuing improvements in ecological modeling.

   

RHESSys estimates ecological processes such as rates of evapotranspiration, hydrological balance, or net primary productivity (Fig. 1). FIRE-BGC is a biogeochemical succession model that uses those estimates to generate the physical structure and species diversity of forests. FIRE- BGC defines homogeneous landscape units (like forest stands) and calculates individual tree growth, death rates, seedling survival, organic matter accumulation, and decomposition both daily and annually. The replacement of one stand of trees by another stand (succession) can be tracked through time as tree demographic processes take place. The role of ecological disturbances, such as large forest fires, has also been integrated (Finney and Ryan 1995; Fig. 3).



Fig. 3. A forest succession model, FIRE-BGC, predicts the occurrence of two major forest fires in the upper McDonald watershed during simulation year 163. Blue colors are the cooler part of the fire, pink the hottest parts. The boundaries and intensity of the potential future fire are calculated by another model, FARSITE, based on such parameters as fuel loads and simulated daily meteorology Courtesy R. Keane, U.S. Forest Service
How Well Do the Models Work?  

We can test model performance by comparing ecosystem processes simulated for present conditions with measurements of the real ecosystem. Climate data for Glacier National Park's Lake McDonald basin have been used to drive a RHESSys simulation for the same period when numerous field measurements were being taken. For instance, thousands of snow depth measurements were made to verify the RHESSys estimates of snowpack distribution and moisture content. This test was critical because snowpack provides significant moisture during summer months for many ecosystem functions. Automated weather stations were placed on remote mountain slopes to confirm that the microclimate model was making reasonable predictions of climatic variables (Fig. 4). Many other ecosystem measurements were taken, including stream discharge volume and timing. Figure 5 shows the relation between the streamflow that actually occurred and what RHESSys predicted would occur (Comanor et al. 1997). Other comparisons of modeled and measured ecosystem phenomena showed similar relationships.



Fig. 4. Research scientists from the U.S. Geological Survey and National Park Service at an automated climate station above Lake McDonald, Glacier National Park. Data collected from this station will help verify the meteorological estimates generated by a submodel of RHESSys. Courtesy F. Klasner, USGS

RHESSys is able to reasonably simulate the underlying dynamics that drive ecological changes in Lake McDonald basin. These simulations indicate that the Glacier National Park landscape is dynamic and will change over time even with a stable climate. We cannot assume, however, that the climate is not changing.

Fig. 5. A comparison of the observed and simulated outflow (meters per second) from the upper Lake McDonald watershed, Glacier National Park, 1993. Simulated outflows were calculated by RHESSys.
© J. White, University of Montana
Predicting Climate Change Effects  

Forecasting future climates is both difficult and controversial. The value of models like RHESSys is that a variety of climate scenarios can be used to simulate the range of outcomes for Glacier National Park. As climate models improve, RHESSys can use the new forecasts to identify increasingly probable ecosystem changes.

   

The simulation for the Lake McDonald basin in Figure 6 is based on the assumption that the regional climate will be 0.5°C warmer and will have a 30% increase in annual precipitation 50 years from now. Predicted changes in vegetation include expanded cedar­hemlock forests, making Glacier National Park resemble the wetter forests of the Pacific Northwest. A different climate scenario suggests Glacier National Park stream temperatures will rise, especially in late summer, because of changes in streamflow timing and volume. Such changes would affect temperature-sensitive aquatic organisms; for example, different species live in each section of stream as the water becomes colder with increasing elevation. Predictable shifts of these organisms upstream and to higher elevations will occur under the RHESSys-projected changes.



Fig. 6. Estimated changes in vegetation cover by 2050 for the Lake McDonald watershed, Glacier National Park, following climate change. Changes were estimated by RHESSys by using a climate change scenario of a 30% increase in annual precipitation and a 0.5°C increase in average annual temperature. © J. White and S. Running, University of Montana

RHESSys accurately predicted snowfall distribution at Glacier National Park, identifying areas where snow never melted completely before the onset of the next winter. These are the same areas that currently support glaciers (White and White 1994). These glaciers have been steadily receding during this century (Fig. 7), and RHESSys suggests that future snowpacks will not survive through summer to nourish glaciers. A separate modeling effort extrapolated glacial melt rates into the future by using different scenarios for global warming (Hall 1994). In this model, whether current warming trends continue or are accelerated by increasing atmospheric carbon dioxide, glaciers will not exist in Glacier National Park by 2030.



Fig. 7. A geographic information system representation of glacier shrinkage from 1850 to 1993 in Glacier National Park. The Blackfeet­Jackson glaciers are in the center. The yellow areas reflect the current area of each glacier; other colors represent the extent of the glaciers at various times in the past. Courtesy C. Key, USGS and R. Menicke, National Park Service
Future Models  

Ecosystem modeling systematically organizes our current ecological knowledge to provide quantitative estimates we can test in the field. Once underlying ecological principles have been affirmed in the model by successful field tests, we can use a model to glimpse the future and help prepare for it. Like climate change models, simulation systems like RHESSys and FIRE-BGC will be continually improved. It seems clear, however, that these ecosystem models will challenge us to think more specifically about the effects of climate change and will be an essential tool for understanding the future status and trends of our biological resources.

   
  Author
Daniel B. Fagre
U.S. Geological Survey
Biological Resources Division
Glacier Field Station Science Center
Glacier National Park
West Glacier, Montana 59936

1996 2050

   

References


Home