The U.S. Geological Survey’s Land Change Monitoring, Assessment, and Projection (LCMAP) initiative is a new end-to-end capability to continuously track and characterize changes in land cover, use, ...and condition to better support research and applications relevant to resource management and environmental change. Among the LCMAP product suite are annual land cover maps that will be available to the public. This paper describes an approach to optimize the selection of training and auxiliary data for deriving the thematic land cover maps based on all available clear observations from Landsats 4–8. Training data were selected from map products of the U.S. Geological Survey’s Land Cover Trends project. The Random Forest classifier was applied for different classification scenarios based on the Continuous Change Detection and Classification (CCDC) algorithm. We found that extracting training data proportionally to the occurrence of land cover classes was superior to an equal distribution of training data per class, and suggest using a total of 20,000 training pixels to classify an area about the size of a Landsat scene. The problem of unbalanced training data was alleviated by extracting a minimum of 600 training pixels and a maximum of 8000 training pixels per class. We additionally explored removing outliers contained within the training data based on their spectral and spatial criteria, but observed no significant improvement in classification results. We also tested the importance of different types of auxiliary data that were available for the conterminous United States, including: (a) five variables used by the National Land Cover Database, (b) three variables from the cloud screening “Function of mask” (Fmask) statistics, and (c) two variables from the change detection results of CCDC. We found that auxiliary variables such as a Digital Elevation Model and its derivatives (aspect, position index, and slope), potential wetland index, water probability, snow probability, and cloud probability improved the accuracy of land cover classification. Compared to the original strategy of the CCDC algorithm (500 pixels per class), the use of the optimal strategy improved the classification accuracies substantially (15-percentage point increase in overall accuracy and 4-percentage point increase in minimum accuracy).
Landsat data have enabled continuous global monitoring of both human-caused and other land cover disturbances since 1972. Recently degraded performance and intermittent service of the Landsat 7 and ...Landsat 5 sensors, respectively, have raised concerns about the condition of global Earth observation programs. However, Landsat imagery is still useful for landscape change detection and this capability should continue into the foreseeable future.
In order to understand the magnitude, direction, and geographic distribution of land-use changes, we evaluated land-use trends in U.S. counties during the latter half of the 20th century. Our paper ...synthesizes the dominant spatial and temporal trends in population, agriculture, and urbanized land uses, using a variety of data sources and an ecoregion classification as a frame of reference. A combination of increasing attractiveness of nonmetropolitan areas in the period 1970-2000, decreasing household size, and decreasing density of settlement has resulted in important trends in the patterns of developed land. By 2000, the area of low-density, exurban development beyond the urban fringe occupied nearly 15 times the area of higher density urbanized development. Efficiency gains, mechanization, and agglomeration of agricultural concerns has resulted in data that show cropland area to be stable throughout the Corn Belt and parts of the West between 1950 and 2000, but decreasing by about 22% east of the Mississippi River. We use a regional case study of the Mid-Atlantic and Southeastern regions to focus in more detail on the land-cover changes resulting from these dynamics. Dominating were land-cover changes associated with the timber practices in the forested plains ecoregions and urbanization in the piedmont ecoregions. Appalachian ecoregions show the slowest rates of land-cover change. The dominant trends of tremendous exurban growth, throughout the United States, and conversion and abandonment of agricultural lands, especially in the eastern United States, have important implications because they affect large areas of the country, the functioning of ecological systems, and the potential for restoration.
Vegetation phenological phenomena are closely related to seasonal dynamics of the lower atmosphere and are therefore important elements in global models and vegetation monitoring. Normalized ...difference vegetation index (NDVI) data derived from the National Oceanic and Atmospheric Administration's Advanced Very High Resolution Radiometer (AVHRR) satellite sensor offer a means of efficiently and objectively evaluating phenological characteristics over large areas. Twelve metrics linked to key phenological events were computed based on time-series NDVI data collected from 1989 to 1992 over the conterminous United States. These measures include the onset of greenness, time of peak NDVI, maximum NDVI, rate of greenup, rate of senescence, and integrated NDVI. Measures of central tendency and variability of the measures were computed and analyzed for various land cover types. Results from the analysis showed strong coincidence between the satellite-derived metrics and predicted phenological characteristics. In particular, the metrics identified interannual variability of spring wheat in North Dakota, characterized the phenology of four types of grasslands, and established the phenological consistency of deciduous and coniferous forests. These results have implications for large-area land cover mapping and monitoring. The utility of remotely sensed data as input to vegetation mapping is demonstrated by showing the distinct phenology of several land cover types. More stable information contained in ancillary data should be incorporated into the mapping process, particularly in areas with high phenological variability. In a regional or global monitoring system, an increase in variability in a region may serve as a signal to perform more detailed land cover analysis with higher resolution imagery.
Landscape pattern and composition metrics are potential indicators for broad-scale monitoring of change and for relating change to human and ecological processes. We used a probability sample of ...20-km x 20-km sampling blocks to characterize landscape composition and pattern in five US ecoregions: the Middle Atlantic Coastal Plain, Southeastern Plains, Northern Piedmont, Piedmont, and Blue Ridge Mountains. Land use/and cover (LULC) data for five dates between 1972 and 2000 were obtained for each sample block. Analyses focused on quantifying trends in selected landscape pattern metrics by ecoregion and comparing trends in land cover proportions and pattern metrics among ecoregions. Repeated measures analysis of the landscape pattern documented a statistically significant trend in all five ecoregions towards a more fine-grained landscape from the early 1970s through 2000. The ecologically important forest cover class also became more fine-grained with time (i.e., more numerous and smaller forest patches). Trends in LULC, forest edge, and forest percent like adjacencies differed among ecoregions. These results suggest that ecoregions provide a geographically coherent way to regionalize the story of national land use and land cover change in the United States. This study provides new information on LULC change in the southeast United States. Previous studies of the region from the 1930s to the 1980s showed a decrease in landscape fragmentation and an increase in percent forest, while this study showed an increase in forest fragmentation and a loss of forest cover.
Satellite-derived contemporary land-cover land-use (LCLU) and albedo data and modeled future LCLU are used to study the impact of LCLU change from 2000 to 2050 on surface albedo and radiative forcing ...for 19 ecoregions in the eastern United States. The modeled 2000–2050 LCLU changes indicate a future decrease in both agriculture and forested land and an increase in developed land that induces ecoregion radiative forcings ranging from −0.175 to 0.432 W m⁻² driven predominately by differences in the area and type of LCLU change. At the regional scale, these projected LCLU changes induce a net negative albedo decrease (−0.001) and a regional positive radiative forcing of 0.112 W m⁻². This overall positive forcing (i.e., warming) is almost 4 times greater than that estimated for documented 1973–2000 LCLU albedo change published in a previous study using the same methods.
Quantifying carbon dynamics over large areas is frequently hindered by the lack of consistent, high-quality, spatially explicit land use and land cover change databases and appropriate modeling ...techniques. In this paper, we present a generic approach to address some of these challenges. Land cover change information in the Southeastern Plains ecoregion was derived from Landsat data acquired in 1973, 1980, 1986, 1992, and 2000 within 11 randomly located 20-km x 20-km sample blocks. Carbon dynamics within each of the sample blocks was simulated using the General Ensemble Biogeochemical Modeling System (GEMS), capable of assimilating the variances and covariance of major input variables into simulations using an ensemble approach. Results indicate that urban and forest areas have been increasing, whereas agricultural land has been decreasing since 1973. Forest clear-cutting activity has intensified, more than doubling from 1973 to 2000. The Southeastern Plains has been acting as a carbon sink since 1973, with an average rate of 0.89 Mg C/ha/yr. Biomass, soil organic carbon (SOC), and harvested materials account for 56%, 34%, and 10% of the sink, respectively. However, the sink has declined continuously during the same period owing to forest aging in the northern part of the ecoregion and increased forest clear-cutting activities in the south. The relative contributions to the sink from SOC and harvested materials have increased, implying that these components deserve more study in the future. The methods developed here can be used to quantify the impacts of human management activities on the carbon cycle at landscape to global scales.
Strategies for mitigating the global greenhouse effect must account for soil organic carbon (SOC) dynamics at both spatial and temporal scales, which is usually challenging owing to limitations in ...data and approach. This study was conducted to characterize the SOC dynamics associated with land use change history in the northwestern Great Plains ecoregion. A sampling framework (40 sample blocks of 10 × 10 km2 randomly located in the ecoregion) and the General Ensemble Biogeochemical Modeling System (GEMS) were used to quantify the spatial and temporal variability in the SOC stock from 1972 to 2001. Results indicate that C source and sink areas coexisted within the ecoregion, and the SOC stock in the upper 20‐cm depth increased by 3.93 Mg ha−1 over the 29 years. About 17.5% of the area was evaluated as a C source at 122 kg C ha−1 yr−1. The spatial variability of SOC stock was attributed to the dynamics of both slow and passive fractions, while the temporal variation depended on the slow fraction only. The SOC change at the block scale was positively related to either grassland proportion or negatively related to cropland proportion. We concluded that the slow C pool determined whether soils behaved as sources or sinks of atmospheric CO2, but the strength depended on antecedent SOC contents, land cover type, and land use change history in the ecoregion.
Land cover change is one of the key driving forces for ecosystem carbon (C) dynamics. We present an approach for using sequential remotely sensed land cover observations and a biogeochemical model to ...estimate contemporary and future ecosystem carbon trends. We applied the General Ensemble Biogeochemical Modelling System (GEMS) for the Laurentian Plains and Hills ecoregion in the northeastern United States for the period of 1975–2025. The land cover changes, especially forest stand-replacing events, were detected on 30 randomly located 10-km by 10-km sample blocks, and were assimilated by GEMS for biogeochemical simulations. In GEMS, each unique combination of major controlling variables (including land cover change history) forms a geo-referenced simulation unit. For a forest simulation unit, a Monte Carlo process is used to determine forest type, forest age, forest biomass, and soil C, based on the Forest Inventory and Analysis (FIA) data and the U.S. General Soil Map (STATSGO) data. Ensemble simulations are performed for each simulation unit to incorporate input data uncertainty. Results show that on average forests of the Laurentian Plains and Hills ecoregion have been sequestrating 4.2
Tg C (1
teragram
=
10
12
gram) per year, including 1.9
Tg C removed from the ecosystem as the consequences of land cover change.
The United States has a highly varied landscape because of wide-ranging differences in combinations of climatic, geologic, edaphic, hydrologic, vegetative, and human management (land use) factors. ...Land uses are dynamic, with the types and rates of change dependent on a host of variables, including land accessibility, economic considerations, and the internal increase and movement of the human population. There is a convergence of evidence that ecoregions are very useful for organizing, interpreting, and reporting information about land-use dynamics. Ecoregion boundaries correspond well with patterns of land cover, urban settlement, agricultural variables, and resource-based industries. We implemented an ecoregion framework to document trends in contemporary land-cover and land-use dynamics over the conterminous United States from 1973 to 2000. Examples of results from six eastern ecoregions show that the relative abundance, grain of pattern, and human alteration of land-cover types organize well by ecoregion and that these characteristics of change, themselves, change through time.