Land Use and Water Quality:
Connecting Ecology and Economics

By Merrill Leffler

As land use goes, so goes the health of Chesapeake Bay - and, for that matter, the health of estuarine and near-shore waters throughout the world.

Land in the Bay watershed, as it has been progressively cleared for agriculture and urban development, has become a conduit for eroding soils and contaminants that flow directly into streams and rivers or indirectly through groundwater seepage that leads to the Chesapeake. An estimate put forward by the Maryland Department of the Environment several years ago held that for every acre of forest cleared for development, some 240 to 420 pounds of sediment enter the Bay.

The impacts of this runoff on the Chesapeake have been as various as they are many - loss of underwater grasses and habitat for fish, depletion of oxygen and changes in food webs are only a few examples.


Researchers have begun to detail just how different land uses affect soil erosion and nutrient transport to estuaries like the Chesapeake.

Old Problems, New Approaches

That water quality depends on land use is not new news. A major aim of the Chesapeake Bay Program, like aquatic restoration programs around the nation, has been to stop soil erosion and slash nutrient loading in runoff and groundwater.

The newer news, however, is that researchers have begun to detail just how different land uses affect soil erosion and nutrient transport to estuaries like the Chesapeake. In the Rhode River, for example, a sub-estuary of the Bay's western shore, David Correll and his colleagues at the Smithsonian Environmental Research Center have for some years been measuring runoff through forests, pastures and croplands.

"Many studies have been done of nutrient cycling and flow through specific ecosystems," says Correll, "but few have analyzed nutrient flows through landscapes containing several different kinds of ecosystems." Long-term measurements are critical, since flows can vary enormously from year to year. As Correll pointed out in a recent workshop on Land-Use Effects in the Mid-Atlantic Region, one pasture area on the Rhode River discharged an average of 100 to 200 parts per billion nitrate during 1989-1991; during 1982 and 1992, however, the mean nitrate concentrations discharged from the same pasture were from 1,000 to 1,300 parts per billion. These wide-ranging variations, the result of annual rainfall differences as well as natural differences in similar-use landscapes, could be missed in short-term monitoring.

Correll's work in the Rhode River and other watersheds in the Bay basin is providing the kind of numbers that clearly demonstrate how pastures, crop land and forest stands radically affect soil erosion and nutrient flow into the Bay. Such measurements provide the kind of detail needed to construct mathematical models that can predict impacts of different types of land development on water quality.

Of Maps and Models

Models, says Robert Costanza of the Chesapeake Biological Laboratory, part of the University of Maryland Center for Environmental Science (UMCES), are like maps. "They are abstract representations of complex territory. While no one map or model is right for an entire range of uses," he points out, "their usefulness can best be judged by their ability to help solve the navigational problems you are interested in." What kind of navigational problems? Perhaps the effect of clearing pasture land for residential and business development in a large river system; or the impact on a local creek of clear cutting a forest stand for a new golf course.

The ability to better predict how proposed changes in land use are likely to affect aquatic health could give planners and resource managers new tools for trying to balance economic development, growth and environmental protection. That ability could give a more rational basis for designing zoning regulations or reaching decisions on new roads or waste treatment plants.

Costanza and his colleagues at UMCES and the University of Maryland's Institute for Ecological Economics have been constructing sophisticated computer-based models that can account for the effect of different types of land uses in a large watershed. Such modeling depends on the technology of geographical information systems, or GIS. GIS is a computerized system for digitally storing, manipulating and analyzing "snap shots" of land or water features from maps and, increasingly in recent years, aerial and satellite-based photography. Land features include such information as its use (forest, crop, residential), elevation and slope, and types of soil that can then be displayed in numbers of different ways.

Costanza and Thomas Maxwell have constructed a "spatial" model that can calculate the flow of water, soil and nutrients through a complex mosaic of different land uses. The model first divides the landscape into cells of some 100 acres each; based on GIS information, each cell is then assigned a particular use - for instance, forest, cropland, pasture, residential - together with a corresponding set of ecological characteristics. (The quality of the assigned ecological conditions depends on long-term measurements and analysis, such as those that Dave Correll has been doing in the Rhode River.)

At the heart of such spatial modeling is the ability to simulate the vertical and horizontal movements of water, soil and nutrients. Using specially designed software programs that Costanza, Maxwell and Ed DeBellvue have been developing, the model can calculate the horizontal exchange of flows among adjacent cells and vertical flows above and below the sediment for different types of land use and climatic conditions.

Modeling the Patuxent River Watershed

Over the last several years, Costanza, Maxwell and DeBellvue have applied their efforts to a Patuxent Landscape Model and related models. The Patuxent Model can mathematically mimic the effect of past land-use changes on water quality and then simulate the effects of proposed changes - for example, the clearing of forests or conversion of agricultural lands for commercial development or new roads or residential housing. They have been using GIS data, produced by the Maryland Office of Planning, data that show striking land use changes between the early 1970s and today.

The Patuxent River's headwaters originate in Montgomery and Howard counties and meander about 70 miles through seven counties before emptying into the Chesapeake at Solomon's Island. The 926-square-mile Patuxent watershed is a complex mix of rapidly growing urban and suburban development, agricultural lands and forests - some 46 percent is forested and 32 percent is in farmland. It is a watershed in which population doubled between 1970 and 1990. With this population growth has come the conversion of agricultural lands and forests for housing, businesses, roads, shopping centers. The resulting changes in the flow of water and soil off the land make their impact in degraded water quality, loss of underwater grass, and the smothering of fish and shellfish habitat.

Linking Economics to Ecology

For the most part, ecological models like the Patuxent Landscape Model, while they are used to run different scenarios of land use changes, generally ignore the way human behavior affects changes in land use, says Nancy Bockstael, an environmental economist with the Department of Agricultural and Resource Economics at the University of Maryland College Park. Ecologists, she says, might analyze the effects of changes in land by comparing the watershed in 1980 and 1995, or they can assume changes that are likely to occur in the future and run their model to see what the impact will be. Those models, however, are not designed to analyze behavioral factors that influence land use change - they impose human behavioral change hypothetically.

As Robert Gardner, a landscape ecologist at the UMCES Appalachian Research Lab, points out, economic factors "are often a dominant force affecting landscape change." At the same time, "the pattern of resources in a landscape often dominates economic development." Ecologists, he says, often ignore the connections between economics and ecology.

With recent support from the Environmental Protection Agency, Bockstael and Costanza have teamed up to connect ecological modeling with economic modeling. Perhaps the most important contribution of the economists, says Bockstael, is to model human land use conversion and how it is related to both the ecological and economic features of the landscape.

Working with Ivar Strand and Jackie Geoghegan, environmental economists at College Park, Bockstael is trying to develop a model that will "predict future land use changes in a specific area given such information as its history, zoning and other land use restrictions, features of the surrounding landscape, and its level of regional economic activity." With such information, she says, "we can predict the probabilities that a parcel of land with certain characteristics will stay in its present land use or convert to alternative uses." For example, scenarios can be developed which examine how different zoning strategies can affect the probability of land use conversions.

Such modeling can help us in other ways, says Bockstael. Suppose nutrient management became mandatory for all agricultural lands - "such a policy could drive more land out of farming, thus increasing the probability of conversion to an alternative use," she conjectures. The economic model could be used to predict the probabilities of such land conversion; it would then drive the ecology model to assess the affects of that land conversion on water quality. It could be, she points out, that alternative uses might have a more negative impact on aquatic health.

Integrating the ecological and economic models is no simple feat. To begin with, says Jackie Geoghegan, the Patuxent Land Model operates on a time scale of a day or less. "The economic models," she points out, "work on an annual basis. This means that the timing of the exchange of information and these different time scales must be carefully thought out." There are also issues of geographical scale. "The ecosystem model divides the study area into cells of some 100 acres each, but for land use conversion decisions, this is much too large relative to individual parcel size," she says. The models are being estimated independently and will eventually be linked in the simulation stage.

The Future

Policies designed to protect the health of coastal waters or other ecosystems are often based on economics, politics and best guesses about ecological impacts of changes in land use. New modeling approaches aim at giving a firmer ground to best guesses that could influence future development.

A major conclusion of the 1991 Governor's Commission on Growth in the Chesapeake Bay Region was the key role of land use to Maryland's environmental future: "How people use Maryland's land determines the fate of Chesapeake Bay." The Commission recommended that development be directed away from farms, forests and other important environmental resources and proposed concentrated development as opposed to sprawl.

According to Ed Rissee of Synergy Planning, Inc., for example, if Fairfax, Virginia (with its suburban sprawl development), were settled like the planned community Reston, Virginia (with its medium density development), two-thirds of Fairfax would likely still be open land available for recreation, agriculture or future planned development. Also, there would have been far less need for extensive roads, new schools or other infrastructures, a potentially large savings for taxpayers.

What are the specific effects of such different types of development on water quality? What are the effects on runoff? On nutrient loading? On fisheries? It is such questions that new developments in modeling may be able to help answer. For planners and policy makers, such capabilities may make it possible to forecast the effects of development and to make better tradeoffs among politics, economics and the environment - achieving a balance that will ultimately save money and help protect the health of the Chesapeake Bay.

Top of Page

		Contents of this issue	Other Issues