Land Cover Change in the Eastern United States

By Thomas R. Loveland ¹ and William Acevedo ¹

Introduction

Land change studies attempt to explain (1) where change is occurring, (2) what land cover types are changing, (3) the types of transformation occurring, (4) the rates or amounts of land change, and (5) the driving forces and proximate causes of change. The ultimate reasons for studying these change characteristics are to understand land change trends, evaluate and manage the consequences of change, and define future scenarios of change. Challenging issues are associated with measuring, gathering, and documenting the characteristics of land change.

The U.S. Government has many programs that address land change. While excellent programs for inventorying forest (Gillespie, 1999) and agricultural land uses (Nusser and Goebel, 1997) exist, because of design, definitions, and geographic coverage of each inventory program, the sum of these programs cannot be combined to provide a comprehensive picture of U.S. land change. This makes it very difficult to draw conclusions and assess the consequences of contemporary U.S. land change.

To provide estimates of land cover and land use change rates for the United States, without regard to land ownership or land use and land cover type, the U.S. Geological Survey (USGS), with support from the U.S. Environmental Protection Agency (EPA), initiated a study of the rates, causes, and consequences of contemporary land use and land cover change. The study uses a probability sampling approach, along with 1973 to 2000 satellite imagery and supporting aerial photography, to measure National land change on an ecoregion-by-ecoregion basis. The ecoregions represent areas with similar biotic and abiotic characteristics, which collectively represent the resource potential and probable responses to natural and anthropogenic disturbances. The defining ecoregion characteristics (e.g., climate, topography, geology, soils, natural vegetation) determine the types of land cover that can exist within an area and influence the range of land use practices that are possible. Estimates of change are based on the interpretation of five dates of Landsat Multispectral Scanner, Thematic Mapper, and Enhanced Thematic Mapper Plus data (nominally 1973, 1980, 1986, 1992, and 2000) for a set of 10- or 20-km² sample blocks randomly selected for each ecoregion. The goal is to provide estimates within one percent of the actual land change at an 85-percent confidence level. This approach—explained in detail in Loveland and others (2002), Sohl and others (2004), and Gallant and others (2004)—provides clear evidence of the geographic variability in ecoregional land change rates, types, and causes. This report summarizes aggregate change in the Eastern United States and specific characteristics of change occurring in each of 20 Eastern U.S. ecoregions.

Regional Synthesis

As part of a national assessment of U.S. land change, the USGS recently completed an analysis of 20 Eastern U.S. ecoregions (fig. 1). The 20 ecoregions span 1,650,930 km², as defined by the EPA (Omernik, 1987; EPA, 1999).

Figure 1. Land cover of the 20 Eastern U.S. ecoregions comprising the “forested east.” Click on image to enlarge

Land cover percentages for all land cover types for all five periods (1973, 1980, 1986, 1992, and 2000) are presented in table 1. The majority land cover of the eastern ecoregions in 2000 was forest (52.4 percent of the region); however, the amount of forest cover has declined since 1973. Since 1973, agriculture (21.6 percent of the region) has also declined, while developed land, primarily related to urban growth, has increased (10.6 percent of the region). Wetlands (7.5 percent) and water (4.0 percent) have stayed relatively consistent in area over the 1973 to 2000 period.

Table 1. Percentages of general land cover types for 1973, 1980, 1986, 1992, and 2000 [The overall area for the 20 ecoregions is 1,650,930 km²]

The average overall amount of Eastern U.S. land change between 1973 and 2000 was 12.5 percent, meaning that 207,000 km² of the 1.65 million km² changed one or more times. However, the average amount masks the geographic variability of change, which ranges from a low of 2.0 percent in the Blue Ridge ecoregion to a high of 24.9 percent in the Southern Coastal Plain ecoregion (fig. 2). The highest amounts of change were generally in ecoregions with active timber harvesting, while the lowest amounts of change were in ecoregions where urbanization was the leading land change (table 2). In the southeastern ecoregions, forest change was dominant. The average amount of change in the seven southeastern ecoregions was 18.9 percent. In the Appalachian ecoregions, where change was far more heterogeneous, the average amount of change was 5.8 percent. In the Megalopolis ecoregions, where urban development dominated land change, the average amount of change was 5.9 percent. The average amount of change in the Northeast was 8.2 percent and was associated with forest land use.

Table 2. Overall change rates by ecoregion [Ecoregions and change rates are in regional groups]

Figure 2. Overall spatial change from 1973 to 2000 for all Eastern U.S. ecoregions. The entire bar shows the overall spatial change, while the gradients indicate the percent of ecoregion area that changed during one or multiple periods. Ecoregions with larger numbers and amounts of change typically have active cyclic forest harvesting, while ecoregions with low amounts of multiple changes are usually characterized by increased urban development. Click on image to enlarge

Summary of Eastern U.S. Land Change

There is no single story of Eastern U.S. land change. Instead, there are 20 distinct stories in the 20 ecoregions. Generally, most change was associated with forest harvesting and regrowth, agricultural abandonment, and development. However, in selected ecoregions with unique natural resources, other cover types corresponding to those unique resources were the key story. For example, coal-bearing geological formations in the Central Appalachians, Southwestern Appalachians, and Western Allegheny Plateau and phosphate reserves in the Southern Coastal Plain were important contributors to land change.

Both net and gross change statistics need to be taken into account when considering land change amounts. For example, the net change in a given land cover category results from the gains and losses in that cover type between two periods, such as 1973 to 2000. Net change represents the difference in land cover between specific points in time. On the other hand, gross change represents the total area modified between two periods (e.g., increases in forest cover and decreases in forest cover) and is important because it provides a clear indication of the overall amount of change activity that affected specific land cover sectors. Gross change provides clear evidence of the overall amount of land change experienced rather than the actual availability of specific land cover.

Table 3. Gross and net changes in land cover classes since 1973

Both net change and gross change statistics for the 10 land cover types are found in table 3. The top five gross changes were as follows:

• Forest (125,473 km²)
• Mechanically disturbed land (77,762 km²)
• Grassland/Shrubland (48,574 km²)
• Agriculture (43,642 km²)
• Developed land (37,398 km²)

On the other hand, when considering net changes that occurred over the 27-year period, the top five land change rankings (highest to lowest) were as follows:

• Forest (–36,946 km²)
• Developed land (36,701 km²)
• Agriculture (–25,022 km²)
• Mechanically disturbed land (20,271 km²)
• Grassland/Shrubland (8,149 km²)

A summary of net change in area per land cover type is found in table 4. A discussion of the land changes follows.

Table 4. Net changes in land cover for the 20 Eastern U.S. ecoregions

Forest Change

The forest sector changed the most. Between 1973 and 2000, 105,437 km² of land were converted to forest cover, while 142,480 km² of forest were converted to non-forest cover. The result was a net loss of 37,044 km² of forest cover, amounting to a 1.5 percent reduction in eastern forest area. This net figure masks overall changes in forest cover. The gross change in forest cover was 125,570 km² of land. This means that 7.6 percent of the total Eastern U.S. area experienced a conversion either to or from forest cover.

Timber production is very important throughout the East. Southeastern ecoregions, especially the Southeastern Plains, Southern Coastal Plain, Middle Atlantic Coastal Plain, and Mississippi Valley Loess Plains, contain vast expanses of industrial-scale pine plantations. These ecoregions have a climate advantage that permits harvest cycles of 20 to 25 years for Pinus taeda (i.e., loblolly pine) and other southern pine species (Gresham, 2002). Large-scale timber production is also important in the Northeast, especially in the Northeastern Highlands and to a lesser extent in the Laurentian Plains and Hills in Maine.

Only the Mississippi Valley Loess Plains ecoregion gained forest cover (0.9 percent increase), while all other ecoregions experienced a reduction. The Piedmont and Southern Coastal Plain ecoregions experienced the highest relative loss of forest (–4.8 percent and –4.2 percent, respectively).

The primary forest conversion was to mechanically disturbed, clear-cut land and forest land recently harvested. More than 31,000 km² of forest changed to this transitional category (fig. 3). This suggests that there is a very dynamic and cyclic nature to eastern forest change. Southern pine forest lands when harvested are classified as mechanically disturbed (e.g., clear cut). The cleared land typically remains in a mechanically disturbed state for a brief period and is then replanted into trees; however, in some cases, the land is left for natural regeneration. In much of the South, we observed that planted pines can reach 2 m within two years. Thus, the forest to mechanically disturbed transition is indicative of extensive timber production in the South. On the other hand, in the northern ecoregions and selected marginal sites in other ecoregions, slower growth rates result in a longer forest cutting and regrowth cycle, so the typical transition is from forest to mechanically disturbed to grassland/shrubland to forest.

Figure 3. Net gains and losses associated with forest cover change. Land cover types above the zero line indicate lands converted to forest, and cover types below the zero line indicate lands converted from forest. Click on image to enlarge

Figure 3 shows that nearly 19,000 km² of forest were converted to developed cover. On the other hand, more than 8,000 km² of agricultural land were either converted to forest (e.g., Southeastern Plains, Mississippi Valley Loess Plains) or abandoned, resulting in afforestation (e.g., Northern Appalachian Plateau and Uplands and Eastern Great Lakes and Hudson Lowlands). Most agriculture to forest conversion was motivated by the relative difference between timber and agricultural profitability. After 1986, however, conversion from agriculture to forest and subsequent gains in forest area were often caused by the Conservation Reserve Program.

Forest land use (i.e., land managed for the production of forest products or maintained as woody vegetation for indirect benefits such as protection of watersheds or for recreational purposes) declined less rapidly than forest cover (i.e., land covered with trees). Note that harvested forest where trees may be replanted or regenerated can be classified as forest use, even though no trees are present. While harvested areas are generally replanted to timber, the cycles of forest use have caused a net decline in the total area of forest cover. Consider that 78 percent of the mechanically disturbed land and 73 percent of eastern grassland/shrubland cover transitions back to forest, implying that the land is continuing to be used for forest production. If those portions of mechanically disturbed land and grassland/shrubland revert to forest, then the net loss of forest land use in the East is 15,219 km².

Developed Land Change

Developed land includes the built-up surfaces of the East. Most are associated with urban and exurban growth, but this category also includes transportation systems and many other dispersed built-up lands. The net increase in developed land (36,761 km²) was almost equal to the loss of forest. This amounts to a 9.4 percent net increase in urban lands. The gross change in developed lands was fairly similar (37,427 km²) due to the unidirectional nature of urban growth. In comparison, between 1970 and 2000, population in the counties corresponding to the 20 eastern ecoregions increased by 32.6 percent, from 93.3 million to 123.8 million people.

Ecoregions with the largest increase in developed land (defined as the percent of the ecoregion converted to developed cover) were the Southern Coastal Plain (6.2 percent), Atlantic Coastal Pine Barrens (4.7 percent), Northern Piedmont (4.6 percent), Piedmont (4.5 percent), Mississippi Valley Loess Plains (4.1 percent), and Northeastern Coastal Zone (4.0 percent). See figure 4 for a map of the net changes in developed land for all ecoregions.

Figure 4. Maps showing the net change in each of the 20 Eastern U.S. ecoregions for the developed, agriculture, forest, and mechanically-disturbed cover types. Click on image to enlarge

Most newly developed land was converted from forest (18,982 km²) and agricultural land (11,685 km²) (fig. 5). Forest to developed conversion was more common in the northeast ecoregions where agricultural land preservation is important. For example, in the Northeastern Coastal Zone, forest was converted to developed land by a 4-to-1 margin over agricultural land, and a 10-to-1 margin was measured in the Northeastern Highlands. On the other hand, between 1973 and 2000, in the Mississippi Valley Loess Plains, development occurred on agricultural land by a nearly 3-to-1 margin, and in the Atlantic Coastal Pine Barrens and Northern Piedmont ecoregions, agricultural land was preferred by a 2-to-1 margin. It is important, however, to recognize that changes over the full 27 years often masked the variability and specific margins of conversion that occurred between each specific time period.

Figure 5. Net gains and losses associated with developed cover change. Land cover types above the zero line indicate lands converted to developed cover types. No development losses occurred. Click on image to enlarge

Developed land increased at an accelerating rate over the 27-year period. The percentage of increased urban land grew every period with average annual developed land increases of 0.06 percent, 0.07 percent, 0.09 percent, and 0.10 percent for each period between 1973 and 2000. While the overall percent gains in developed land are relatively small in relation to the overall area, the increasing rate of development may alter hydrological processes, contribute to urban heat island effects, threaten biodiversity, and result in many other environmental consequences.

Agricultural Change

Agricultural land cover declined almost everywhere, with 2.4 percent net loss (24,891 km²) between 1973 and 2000. Gross agricultural change was 43,676 km². The only ecoregion that did not experience a loss of agricultural land was the Southern Florida Coastal Plain. This ecoregion experienced a significant increase in agricultural cover between 1973 and 1980 because of the need to replace banned sugar cane from Cuba (Walker and Solecki, 2004). Even in this ecoregion, however, agricultural cover has declined since 1980, and the net gain over the study period was less than 0.1 percent.

Agricultural conversion to developed land was the most common conversion in the Megalopolis ecoregions (11,685 km²) (fig. 6). These conversions were generally important throughout the Eastern United States, although conversion to forest was more common in the Southeast where timber production is high. Ecoregions with significant agriculture to developed land conversions were the Northern Piedmont and Atlantic Coastal Pine Barrens, where total agricultural land losses were 3.4 percent and 2.9 percent, respectively (See figure 4 for a map of net changes in agricultural land for all eastern ecoregions). The specific factors behind the higher rates of agriculture to developed land conversions must still be determined for each specific ecoregion, but factors likely include the lower costs to develop the relatively level sites with better engineering soil properties.

Figure 6. Net gains and losses associated with agricultural land change. Land cover types above the zero line indicate lands converted to agriculture, and cover types below the zero line indicate lands converted from agricultural cover. Click on image to enlarge

Significant conversions of agricultural land to forest cover occurred in the Mississippi Valley Loess Plains and Southeastern Plains, where more than 8,000 km² were converted from agriculture to forest cover. Rates of agricultural loss accelerated in the 1986 to 1992 period, indicating that an important driver of change was the enrollment of marginal croplands in the Conservation Reserve Program.

An additional transformation important in the northeast ecoregions (e.g., in the Northern Appalachian Plateau and Uplands and the Eastern Great Lakes and Hudson Lowlands) is the abandonment of agricultural lands. Nearly 3,500 km² of agricultural land reverted to grassland/shrubland following cropland abandonment. These lands will likely revert to forest.

Transitional Lands

A hidden dimension of Eastern U.S. change involves transitional land cover, i.e., mechanically disturbed, non-mechanically disturbed, and grassland/shrubland. As discussed earlier, harvested forest planned for replanting or regeneration goes through either a mechanically disturbed state, the typical transition in southern ecoregions, or a mechanically disturbed-grassland/shrubland transition, typical in the northern ecoregions. Afforestation associated with abandoned agricultural land also transitions through a grassland/shrubland stage.

At any one time, the amount of grassland/shrubland or mechanically disturbed land was relatively small. For example, in 1980, the amount of grassland/shrubland was 11,565 km², and the amount of mechanically disturbed land was 13,218 km². Between 1973 and 2000, the net change in grassland/shrubland was 8,125 km², and the net change in mechanically disturbed land was 20,242 km². Together, the net gain in transitional land between 1973 and 2000 was 28,367 km², making this the third largest net land cover change. The gross amount of transitional land over the 27-year period was 126,400 km². This makes transitional land the largest gross change that occurred in the 20 eastern ecoregions.

All 20,200 km² of mechanically disturbed land are transitional, with 78 percent destined for forest cover. Most of the remaining mechanically disturbed land reverts to wetlands, and these cases relate primarily to timber harvesting of wooded wetlands in southeastern wetlands, with a smaller percentage transitioning to developed cover. Approximately 73 percent of the grassland/shrubland reverts to forest cover, and most of the remainder is transformed to agricultural cover.

Non-mechanically disturbed land includes changes associated with wildfires, insect infestations, storms, and other similar events. Overall, very few instances of non-mechanical disturbances occurred in the East. Fires in the Southern Coastal Plain and Southern Florida Coastal Plain and tornado damage in the North Central Appalachians were the major examples of non-mechanical change.

Transitional lands represent short- to medium-term disturbances. Mechanically disturbed lands, representing 20,200 km², are particularly prone to erosion and sedimentation of nearby streams. These areas also affect weather and climate variability by increasing surface albedo and alter regional carbon balances due to the removal of significant carbon stocks. They also affect habitat conditions and, ultimately, biodiversity.

Water and Wetlands Change

Water and wetlands covered 4.5 percent and 8.3 percent, respectively, of the 20 eastern ecoregions. Over the 27-year study period, water increased by 1.3 percent, while wetlands declined by 1.7 percent. The Southern Coastal Plain (0.6 percent) and Southern Florida Coastal Plain (0.4 percent) experienced the largest increase in surface water. In both ecoregions, waterfront residential developments resulting from wetlands dredging or other forms of excavation were significant reasons for the increase. In the Piedmont (0.2 percent), the increase was associated with reservoir developments required to meet the needs of the expanding population base. The Piedmont has limited groundwater resources, so municipal and industrial water relies on surface water impoundments (Daniel and Dahlen, 2002).

Approximately one-third of the 20 ecoregions experienced essentially no change in wetlands area. The ecoregions that experienced significant loss of wetlands included the Southern Coastal Plain (–2.4 percent), the Southern Florida Coastal Plain (–1.6 percent), and the Middle Atlantic Coastal Plain (–1.3 percent). All three ecoregions have extensive wetlands. Agricultural development and urbanization contributed to wetlands loss in the Middle Atlantic Coastal Plain and the Southern Florida Coastal Plain, while urban growth was the primary factor in the Southern Coastal Plain.

Mining

Land area being mined remained unchanged across the East between 1973 and 2000. However, this is the result of a balance between ecoregions with expanded mining (e.g., Central Appalachians coal mining expanded to cover an additional 1.5 percent of the ecoregion) and several ecoregions with mining losses (e.g., Southern Coastal Plain mining shrunk by 1.2 percent of the ecoregion due to a reduction of phosphate mining). Declines in coal mining in the Southwestern Appalachians and Western Allegheny Plateau resulted in a loss of 0.9 percent and 0.4 percent, respectively, of their ecoregion areas. It is also notable, though smaller in scale, that rapidly urbanizing ecoregions, such as the Northeastern Coastal Zone, usually had an increase in aggregate mining over time. Exhausted mined lands were typically restored to grassland/shrubland or forest.

Barren Land Change

No significant changes in barren cover were found in the 20 eastern ecoregions.

Summary

Eastern U.S. land change is connected primarily to timber harvesting and urban growth. The climate characteristics of the southern United States provide the foundation for forest activity. Short harvest cycles mean that a substantial amount of land is cycling from forest to cleared conditions (mechanically disturbed) and back to forest. Across the East, agricultural lands are being converted to forest (e.g., in the southeast ecoregions) or reverting to forest (e.g., in the northeast ecoregions) due to agricultural abandonment. Urban expansion, as measured by the increase in developed cover, is expanding and accelerating across the East. All ecoregions are experiencing significant increases in developed land. Agricultural lands are being lost to urban growth and to forestry practices. Also associated with timber harvesting and expanded urban development is the very large amount of transitional, disturbed lands. Timber harvesting and replanting has created a forest to mechanically disturbed transition in the Southeast and a forest to mechanically disturbed to grassland/shrubland transition in the Northeast. Transitional lands represent a significant amount of disturbance that may have adverse consequences for environmental and ecological systems. Other land cover types, such as mining, water, and wetlands, are changing, but the rates of change are modest. The local importance of those changes may be significant and are best understood at the ecoregion level.

The diversity of the mosaic of change in the Eastern United States is lost when viewed as a composite. In the remainder of the report, the status and trends of land change in each of the 20 eastern ecoregions is presented.

¹ Thomas R. Loveland, William Acevedo – U.S. Geological Survey, Center for Earth Observations and Science, Sioux Falls, SD 57198

References