Forests cover about one-third of the United States and range from wildland forests to urban forests. These diverse ecosystems provide a variety of habitats for wildlife; help to cleanse the air and water; supply timber, fuelwood, and other harvested products; serve as places for recreation; help to mitigate the effects of global climate change; and provide other essential goods and environmental services.

Forests are vulnerable to conversion to other land uses. An increasing number of houses and other buildings in and near forests portend growing costs and complications in fire suppression and potential loss of many values derived from forests. Long-term assessment of the condition of forests and of the relationships between forest conditions and socio-economic factors is the key to defining policy questions and actions needed to sustain forest-based services.

In this synthesis, we survey recent trends, determinants, and projections of forestland conversion in the United States. Examples with more detailed treatments, supporting tables, and figures are available in Alig and others (2003, 2004) and Alig and Ahearn (2006). Forestland conversion is a persistent issue for managers and policymakers; for example, a recent position statement concerning loss of forestland by the Society of American Foresters (2004) lists ecological effects, (e.g., effects on water quality and wildlife habitat) and socio-economic effects, (e.g., expansion of the urban-forest interface, reduction of forest recreation opportunities, reduction of long-term timber production possibilities, and loss of open space) as important implications of forest loss. We examine how socio-economic drivers of land use change, such as population totals, and personal income levels have increased substantially since the Second World War and led to changes in forest ecosystem attributes. We summarize determinants of land use changes, focusing on the societal and private trade-offs of retaining land in forests. Our projections reflect population growth that spurs demand for land for developed uses at the same time that demands for some forest products and other forest benefits are increasing. Risk and policy considerations necessitate that creating effective policy in this area will require careful deliberation concerning private and social viewpoints. For example, some forest benefits (such as wildlife habitat and other ecosystem services) can most effectively be produced at scales greater than the individual private parcel scale and because market imperfections can cause some social forest benefits to be undersupplied when this is the case (Kline and others 2004a).

Five categories of significant changes affecting forest area are:

1. afforestation

2. deforestation

3. forest fragmentation

4. forest parcelization

5. increased numbers of structures, such as houses, on forest land

This analysis does not address changes in forest cover type. For an example of a national analysis on this topic, see Alig and Butler (2004). Examination of intensification of land management is illustrated by the 2001 RPA Timber Assessment (Haynes 2003).

(Table: Data bases used) lists supporting major data bases, and (Table: Example studies) lists examples of studies that have examined land base dynamics in the United States. In the United States, millions of acres of land shift uses each year (USDA NRCS 2001) reflecting billions of choices made by individuals, corporations, nongovernmental organizations, and governments. Next, we look at recent trends in those five categories of changes in the land base. Examining historical trends provides guidance for identifying key factors that are likely to influence forest land condition and associated natural resources in future years. Discussion of historical trends is a foundation for considering projected changes.

From 1953 to 1997, a majority (26) of States had a loss in forest area according to periodic surveys by the USDA Forest Service, e.g., Smith and others (2004). Nine States had net losses of at least 1 million acres each, ranging up to 6.3 million acres. In descending order of net loss amount, the States are: Texas, Florida, California, Oklahoma, Louisiana, Washington, Alaska, Missouri, and Minnesota. Seven States had net gains of more than one million acres, ranging up to 4.1 million acres. In descending order of net gain amount, the States are: New York, Ohio, Pennsylvania, West Virginia, Mississippi, Alabama, and Kentucky. The only regions with net gain in forest area were the North, where a relatively large amount of pastureland reverted naturally to forest, and the Intermountain Region, where a large number of acres were reclassified from pasture or rangeland to forest over time.

Note that much of the shift from pastureland or rangeland to forest use is due to reclassification over time. As trees grow, they expand to reach the 10-percent canopy cover used to define forest land, which changes the pastureland classification to forest use. Even though now classified as forest, the land may still be used for grazing. Shifts between grazing land and forest uses are common, and although they are technically considered as shifts into and out of agriculture, they really represent multiple, overlapping uses.

Conversion of forestland by destination

The long-term loss in United States forest area since the early 1950s has been due to a combination of factors, but, in more recent decades, has been primarily due to conversion to urban and developed uses. Deforestation is conversion from forest to nonforest use, and between 1982 and 1997, 23 million acres were deforested on non-Federal land in the United States. Here we concentrate on private lands and secondarily on other non-Federal lands, for which more data are available, (e.g., USDA NRCS 2001). The destination of about half of the converted forest acres was to urban and developed uses Between 1982 and 1997, more than 10 million acres of non-Federal forests were converted to developed uses, an area larger than the combined current forest area of five Northeastern States (Connecticut, Delaware, Maryland, New Jersey, and Rhode Island). In the most recent data remeasurement period (1992 -97), the proportion of deforestation due to urban and developed increased to 55 percent (USDA NRCS 2001).

Net changes (area into forest minus area out of forest) are typically much smaller than total or gross changes (area into forest plus area out of forest). Gross change in area of non-Federal forests in the contiguous 48 States between 1982 and 1997 totaled about 50 million acres (USDA NRCS 2001). The gross change in forest area was 14 times as large as the net change in forest area.

Development spans a broad range of population density associated with settlement patterns, and definitions of development can depend on the data source and the purpose for which the data are analyzed. Two major data sources both show a steady increase in developed uses over recent decades. Estimates from the U.S. Census Bureau extend farthest back in time and show a 130-percent increase in census-defined urban area between 1960 and 2000. Census urban area is comprised of all territory units in urbanized areas and in places of more than 2,500 persons outside of urbanized areas. The Census measure of urbanization labels as ?built-up? some land that is still to some extent available for rural productive uses, thereby probably erring on the side of overgenerous inclusion (Alig and Healy 1987). Although the term ?paved over? has frequently been used to describe urban land, only a small fraction of the land so classified is literally paved.

The other major data source is the National Resource Inventory (NRI) (USDA NRCS 2001), and it covers a shorter period (1982-97). The NRI estimate of United States developed area increased 34 percent between 1982 and 1997, with an acceleration in the 1990s that was more than 50 percent higher than that of the previous 5 years of measurement. Between 1982 and 1997, developed area as a percentage of the total land area in the 48 contiguous States increased from 3.9 percent to 5.2 percent. Forests were the largest individual source of developed land.

One important feature of the NRI data classification in contrast to the Census urban data is the attempt to exclude areas devoted to agricultural crops, forestry, or similar purposes when they are within a parcel or contiguous area that is otherwise built-up. Outside urban areas, the NRI also includes developed land occupied by nonfarm rural built-up uses, (e.g., rural transportation land), which are not included in the Census urban category. Including transportation infrastructure can be important in that new roads open land to development, alter the environment, (e.g., facilitate invasion of certain species), can create congestion, and can degrade the quality of life. Changes in rural land use have historically been and remain connected with changes in motor vehicle use, technology, and policy.

A significant amount of low-density development has been part of the expansion in developed area. Rural America is home to a fifth of the Nation?s people, and rural residential lots tend to be larger than housing lots in urban areas. One factor in the relatively greater increase in rural residential land use is that it is generally land extensive compared with the land-intensive residential use in urban areas. Rural residential lots, although fewer in number than urban lots, tend to be larger, averaging nearly 3 acres per household, compared with less than a half-acre per household in urban residential areas (USDA ERS 2006). Forty-four million acres, 60 percent of all rural residential lands, are in the largest lot-size category, over 10 acres. Rural land in this category is 3 1/2 times as large as the area of urban land in this category. The wide acreage disparity between rural and urban large-lot categories is likely attributable to relative land values?lower land prices in rural areas make large lots more affordable (USDA ERS 2006).

The low-density housing development in rural areas means more people living closer to remaining forestland. A measure added in recent periodic Forest Inventory and Analysis (FIA) surveys conducted by the USDA Forest Service has been the identification of forestlands by rural-urban continuum class. Based on nationwide rural-urban continuum classes (Smith and others 2004), 13 percent of United States forestland now is located in major metropolitan counties, and 17 percent in intermediate and small metropolitan counties and large towns, together making up 30 percent of all United States forestland (Smith and others 2004, p. 47). Between 1997 and 2002, the forest area in major metropolitan areas increased by 5 percent, or more than 5 million acres, as the United States developed area expanded considerably. Consider that for the whole United States, more than one-quarter of counties are currently classified as metropolitan. That compares with less than one-tenth 50 years ago.

Amount of urban land per additional person is higher for non-metropolitan counties. Many Americans prefer to live in less-congested areas and will commute additional minutes or hours to realize their goals, taking advantage of the United States? excellent road system. Moreover, an increasing population of retirees has augmented out-migration from central cities and suburbs to rural areas that offer aesthetic amenities. Natural amenities may be a more important determinant of county-level inmigration than nearness to metropolitan centers or type of local economy (McGranahan 1999).

South

The largest increases in United States developed area between 1982 and 1997 were in the South, a key timber supply region (USDA NRCS 2001). There, the amount of land in urban and other developed uses increased more than 50 percent since the 1960s. The South had one-third of its developed area added during those 15 years, equal to about half of the United States total of developed area added during that period. Between 1982 and 1997, the South had 7 of the 10 States with the largest average annual additions of developed area according to the NRI. The top three?Texas, Florida, and North Carolina?each added more developed area than did the country?s most populous State, California. Over a more recent period, 1992 - 97, 6 of the 10 States that lost the most cropland, forests, and other open spaces to urban development were in the South. These six southern States in descending order of amount converted were: Texas, Georgia, Florida, North Carolina, Tennessee, and South Carolina (USDA NRCS 2001).

In the Southeast, the concentration of development has been in the area of the urban Piedmont Crescent, extending from Richmond to Atlanta. Within this area are the Interstate 85 and Interstate 40 corridors, the backbone of job growth in the Southeast. Many of the smaller cities are adjacent to larger urban areas, resulting in population concentrations in larger metropolitan areas. The urban areas of the Piedmont are likewise expected to witness the fastest growth, whereas the mountains and the Coastal Plain will experience most of their growth in nonmetropolitan areas.

Several factors contribute to expansion of developed area in the South:

1. above average county population growth due in part to climatic factors and attractiveness to immigrants (Glaeser and Shapiro 2001)
2. above average marginal consumption of land per additional resident
3. income growth

The Southern Forest Resource Assessment (Wear and Greis 2002) identified urbanization as one of the primary threats to forests in the region.

North

Areas of urban and developed uses steadily increased in the North since 1982. Between 1992 and 1997, the area of urban and developed area in the Northeast increased from 10.4 to 11.9 percent of the land base. Corresponding increases in the North Central subregion were from 6.7 to 7.3 percent.

The North had about one-third of the total addition to United States developed area between 1982 and 1997. The North had 3 of the 10 States with the largest average annual additions of developed area according to the NRI.

West

The West?Great Plains, Southwest, California, and Pacific Northwest?accounted for less than one-fifth of the total national addition to NRI developed area between 1982 and 1997. However, recent growth in the region has been above the national average. A growing number of ?ranchettes? and large-lot subdivisions characterize housing growth in the Rocky Mountain region, resulting in the highest amount of developed area per additional person between 1992 and 1997 (Alig and others 2004, USDA NRCS 2001).

The largest percentage of change for a major land use in the contiguous three Pacific Coast States (California, Oregon, and Washington) was the 262-percent increase in urban area between 1960 and 1997 (Alig and others 2003). Urban area as a percentage of total land varies notably by State: 5.9 percent for California, 3.2 percent for Washington, and 1.0 percent for Oregon (Vesterby and Krupa 2001). The State of Washington illustrates the importance of migration for regional population growth and the concentration of growth in coastal areas (Alig and White 2007). Between 1990 and 2000, net migration to western Washington was 180 percent of the natural increase (births?deaths). Approximately 3.5 million people (59 percent of Washington State residents) live within 10 miles of coastline (including the Pacific Ocean and sounds).

More than 8 million acres of forestland were converted to agricultural uses between 1982 and 1997 (USDA NRCS 2001). About half of the converted land has gone to pasture use, with the remainder fairly evenly split between crop use and rangeland. Forestland contributed 55 percent of the land that shifted into agriculture from 1982 to 1997, as land continued to be converted from less intensive uses, like forest, to agricultural uses, like cropland and pasture.

As with the conversions of forestland to developed uses, the majority of forest to agricultural conversions was in the Eastern United States and concentrated in the South. The South had the majority of forestland involved in either conversion to agriculture or gained from agriculture. In the South, land is often suitable for multiple land uses, given relatively gentle topography and ease of access.

On net, forestry gained 14.4 million acres from agriculture between 1982 and 1997. Of total land shifting out of agriculture, 22.7 million acres (46 percent) shifted into forest use, with about 17 million acres being former pastureland. Much of the shift from pastureland to forest use is due to reclassification over time. Factors associated with afforestation can vary by region and over time (in the North, for example, some land formerly used for dairy operations has reverted to forest cover). Most of the afforestation across the Nation—forestation either by human or natural forces of nonforestland—has been of a passive nature, e.g., reclassification as forest cover increased primarily through natural succession. However, tree planting has played a role especially in the South where 25 million acres of pine plantations have been established since 1952, (Alig and Butler 2004) mostly on land formerly in pasture and range use.

Land use change can lead to forest fragmentation?the transformation of a contiguous patch of forest into disjunct patches. Forest fragmentation is widely considered to be a primary threat to terrestrial biodiversity (Armsworth and others 2004), and recent analysis of the fragmentation of continental United States forests indicates that it is so pervasive that edge effects potentially influence ecological processes on a majority of forested lands (Riitters and others 2002).

Definitions of forest fragmentation vary and are influenced by the questions or policy issues of interest. One major distinction is between treating fragmentation as a process and treating it as a pattern, (e.g., Alig and others 2000). Here we discuss forest fragmentation as a pattern. Fragmented forests may occur naturally across the landscape (as in the Great Basin, NV), or this pattern may be a result of human activities, resulting in edge, core or interior habitat, and interspersion changes (Butler and others 2004). Forest fragmentation can be quantified spatially using various indices of landscape structure, with different metrics for different scales of analysis and measurements of interest. Although many fragmentation statistics are available, none provide a definitive indicator of landscape fragmentation, only a means for comparing the characteristics and relative degree of fragmentation across landscapes or periods. Numerous biophysical studies have provided snapshots of forest fragmentation, primarily for the East (Table: Data bases used).

Forest parcelization is the subdivision of forest tracts into smaller ownerships. This phenomenon can have profound impacts on the economics of forestry and lead to reduced forest management, even when land is not physically altered. Land ownership can influence forestland management and investment practices. In addition, per unit costs of forest management practices will increase if economies of scale are lost.

Many of the forest-related increases in population density have been on nonindustrial private forestlands (NIPF) , the ownership class most subject historically to land use changes, e.g., Smith and others 2004. Because NIPF owners are aging and have descendants who live farther from the forest and for whom timber management is not a primary objective, dealing with real estate appreciation may be more central to family succession planning now than it was in the past. Critical wildlife habitat is often provided by NIPF ownership,, as in the Pacific Northwest, where lowlands and riparian areas critical to threatened and endangered species are primarily in NIPF ownership (Bettinger and Alig 1996). Family forests are a large component of the NIPF ownership class; the number of family forest owners increased from 9.3 million in 1993 to 10.3 million in 2003, and these owners now control 42 percent of the Nation?s forestland (Butler and Leatherberry 2004).

Recent shifts in the ownership of the most intensively managed forests in the United States could lead to a substantial increase in parcelization. In the sales of large forest properties, there often is a spin-off for real estate development purposes (highest and best use), and overall, the amount of large industrial forest ownership has been reduced materially in a relatively short time. A large share of the forests long held by consolidated forest products companies has recently been sold to institutional investors. Many of these transactions have occurred in the South. Institutional investors currently hold about 8 percent of the investable United States timberland (Wilent 2004). By the end of 2003, the top 10 timberland investment organizations (TIMOs) managed about 9 million acres of United States timberland, and some analysts predicted that TIMOs and other investor groups, (e.g., Real Estate Investment Trusts, or REITs) will purchase another 10 to 15 million acres in the next decade (Wilent 2004 ).

Two main types of investment models are pursued by TIMOs: separate accounts and closed-end funds. Whereas separate accounts tend to be managed for the long term, closed-end accounts are typically held for a more limited time period of 10 to 15 years before being sold off. One estimate is that one-half of all TIMO investments are of the closed-end type (SAF 2004). In terms of forest fragmentation and conversion, it is the closed-end accounts that may exacerbate rates of deforestation. When TIMOs sell land, they pursue the highest value they can receive, which will most likely be for development and real estate. Currently, sales and acquisitions of forest industry (FI) lands continue to be active as market forces, globalization, and consolidation impact the forest sector.

The emergence of timberland holding firms with timber production objectives but no link to processing facilities has created some difficulties in this traditional taxonomy. Lacking processing facilities, these firms would be grouped in the NIPF class. Yet their timber management behavior is more closely akin to that of the integrated firms in FI. Shifts from traditional integrated FI ownership to the TIMO/REIT class were extremely rapid in the late 1990s and early 2000s, and it is likely that the future will see still further decline in traditional, integrated FI ownership.

With a substantial amount of prime United States timberland shifting from being a personal or industrial asset to being a financial one, more frequent turnover in forest ownership may be part of a new era in forest ownership. This warrants increased attention in data collection and land base monitoring because such changes have implications for a broad range of forest-based ecosystem goods and services due to the influence of changing forest ownership patterns on forest conversion, fragmentation, and parcelization.

A significant proportion of forestland undergoing development each year is used for dispersed residential development in fringe suburbs and smaller cities, commonly known as sprawl. Sprawl is characterized by low-density residential and commercial settlements, and increases in housing density on or adjacent to forests can result in changes to the forest?s quality and function and changes in forest investment (e.g., Kline and others 2004). Forestlands are very popular as residential building sites; forests provide homeowners with shade, screening from neighbors, scenic views, wildlife and bird watching opportunities, and often, easy access to forest-based recreation opportunities. New communication and transportation technologies reduce the isolation of remote locations and make possible long-distance commuting and a wide variety of remote work arrangements. These developments effectively reduce the costs associated with living far from cities and towns. Many areas of the United States are experiencing residential growth in the forests and pressure to develop remaining forests.

Wildland-urban interface in the U.S

The colocation of houses and forests (as well as other wildlands) is captured in the national map of the Wildland Urban Interface (WUI), which was created to aid analysis of the national wildland fire situation. The WUI definition that guided creation of this map and analysis is found in the Federal Register (USDA and USDI 2001) and specifies minimum housing density of 1 structure per 40 acres (or 6.17 structures per km²) and either co-location with, or close proximity to, wildland vegetation. National Land Cover Data and Census Bureau housing data are used together to determine where these conditions exist. Two main types of WUI?intermix and interface?-are identified. Intermix exists where housing and wildland vegetation (at least 50 percent of all pixels in the census block are forests, grasslands, or shrublands) coincide. Areas that meet the housing density minimum but where the wildland vegetation is less dense are considered interface if they are within 1.5 miles of extensive wildland vegetation (defined as an area larger than 5 km² or 1,235 acres with >75 percent wildland vegetation). Together, the intermix and interface make up the WUI (Radeloff and others 2005).

In areas where forest conditions, weather, and climate make wildfire possible, the WUI is a zone where the threat of loss due to wildfire is high, because fires can be carried into this zone where they will threaten homes and lives. Consequently, the WUI has high priority for wildfire hazard reduction treatments. The WUI is also the area where wildland fire outreach programs focus their attention. Resource managers and their outreach partners work with communities and property owners to mitigate wildfire hazards and to plan for evacuation and other emergency measures in the event of wildfire.

Across the United States, the 1990s were a period of rapid housing growth, with a net gain of 13.5 million housing units, a rate of 13 percent growth. The WUI was clearly a preferred setting for new housing; overall WUI growth was 22 percent, and intermix growth was 37 percent. The growth patterns for the United States were consistent across the regions, with growth slower in the non-WUI and faster in the intermix WUI. Most of this WUI housing growth took place in areas already designated as WUI in 1990. The growth in WUI area due to new neighborhoods reaching the housing density minimum for WUI was just 1.5 percent nationally, though it expanded much more in the South and North than in the West.

Across the United States as a whole, the distribution of housing units across the high, medium, and low density interface categories changed little over the decade. Growth in the intermix WUI occurred at high, medium, and low densities, with biggest gains in medium-density intermix. This finding is consistent with adding housing units to existing WUI areas (since existing WUI already had at least low-density housing in 1990) at a greater rate than adding new areas to the WUI. However, in the West, housing growth was strongest in high-density intermix.

In the South, housing increased by 18 percent, almost as much as it increased in the Rocky Mountain region and presented an even greater contrast between non-WUI housing growth (9 percent) and WUI housing growth (29 percent). Over 3 million housing units were added to the WUI during this decade, and the WUI expanded to cover 17.1 percent of the land area, a greater share than in any other region.

Housing grew more slowly in the North than in any other region at just 9 percent. However, the intermix WUI gained nearly 1.2 million new housing units, an increase of 21 percent over the decade. The area of the WUI also expanded; by 2000, over 15 percent of the North was WUI.

Housing increased by 23 percent in the Rocky Mountain region more than in any other region. Once again, WUI housing growth was even stronger (37 percent), whereas intermix WUI housing grew by 75 percent. Although the 2000 WUI makes up just 1.4 percent of this region?s land area, it contains 45.7 percent of the housing units in the State.

Housing growth in the West Coast region was 12 percent overall, with over 1 million new WUI housing units, an 18-percent increase in the number of WUI homes. WUI area also expanded from 5.8 to 6.5 percent of the 3-State area. Of the 16.1 million housing units in this region, over a quarter (4.5 million) are located in the interface WUI.

Analyzing housing growth within the WUI classification provides insight about more than the location and density characteristics of recent change; it also indicates the impact of this growth on forests, grasslands, and other wildland vegetation. More than 60 percent of housing units built in the 1990s were constructed in or near wildland vegetation. Although the fire management community originated the WUI concept as an approximation of where values are at risk from wildland fire, the WUI zone is significant for a broader range of ecosystem services. Clean water, timber, recreation, and other services and outputs from undeveloped land are at risk when development encroaches, and encroachment was significant during the 1990s.

The extent of urban forest has grown appreciably in recent decades. As urban lands expand into surrounding areas, natural resources are often affected or displaced. Between 1990 and 2000, most urban expansion in the United States was on forested or agricultural land (Nowak and others. 2005). Urban areas in the United States, as defined in the 2000 census, contain approximately 3.8 billion trees with an average tree canopy cover of 27 percent (Nowak and others 2001). The impact of current urban vegetation on environmental quality nationally is on the order of several billion dollars per year, (e.g., Nowak and Crane 2002; Nowak and others 2006).

Urbanization concentrates people, materials, and energy into relatively small geographical areas to facilitate the functioning of an urban society. Urbanization often degrades local and regional environmental quality as natural landscapes are replaced with anthropogenic materials. Byproducts of urbanization, (e.g., heat, combustion, and chemical emissions) affect the health of local and regional landscapes, as well as the health of people who visit or reside in and near urban areas. Urban vegetation, through its natural functioning, can improve environmental quality and human health in and around urban areas, with benefits including improvements in air and water quality, building energy conservation, cooler air temperatures, and reduction in ultraviolet radiation.

Five types of land base changes (afforestation, deforestation, forest fragmentation, forest parcelization, and increased number of buildings and people on forestland) have significantly altered United States forests over the last half century. Although net changes in total forest area are relatively small from a national perspective, many more forest acres are actually involved in land use changes as the gross amount of change is more than 10 times the net amount. Small net changes in forest cover do not necessarily equate to small net changes/losses in services provided by forests. The gross changes reflect the combined outcome of the five types of land base changes, which often result in spatial rearrangement of land uses and land covers. The South, which now provides more timber harvest than any other region of the country, in particular, has seen many forest-related land use changes due to population growth and economic activity. In addition to deforestation, substantial forest ownership changes in the South include the shifting of prime timberland from a personal or industrial asset to a financial one. Further, many remaining forest acres are affected by addition of houses on them or nearby. For some forests that are converted to urban uses, there are opportunities to manage urban trees to reduce some of the adverse environmental and health effects associated with urbanization. The South also has a relatively large number of afforestation opportunities, including land suitable for biofuels production as part of global change mitigation strategies.