Predicting land cover using GIS, Bayesian and evolutionary algorithm methods

Modelling land cover change from existing land cover maps is a vital requirement for anyone wishing to understand how the landscape may change in the future. In order to test any land cover change model, existing data must be used. However, often it is not known which data should be applied to the problem, or whether relationships exist within and between complex datasets. Here we have developed and tested a model that applied evolutionary processes to Bayesian networks. The model was developed and tested on a dataset containing land cover information and environmental data, in order to show that decisions about which datasets should be used could be made automatically. Bayesian networks are amenable to evolutionary methods as they can be easily described using a binary string to which crossover and mutation operations can be applied. The method, developed to allow comparison with standard Bayesian network development software, was proved capable of carrying out a rapid and effective search of the space of possible networks in order to find an optimal or near-optimal solution for the selection of datasets that have causal links with one another. Comparison of land cover mapping in the North-East of Scotland was made with a commercial Bayesian software package, with the evolutionary method being shown to provide greater flexibility in its ability to adapt to incorporate/utilise available evidence/knowledge and develop effective and accurate network structures, at the cost of requiring additional computer programming skills. The dataset used to develop the models included GIS-based data taken from the Land Cover for Scotland 1988 (LCS88), Land Capability for Forestry (LCF), Land Capability for Agriculture (LCA), the soil map of Scotland and additional climatic variables.     

Assessing the impact of urbanization on regional net primary productivity in Jiangyin County, China

Urbanization is one of the most important aspects of global change. The process of urbanization has a significant impact on the terrestrial ecosystem carbon cycle. The Yangtze Delta region has one of the highest rates of urbanization in China. In this study, carried out in Jiangyin County as a representative region within the Yangtze Delta, land use and land cover changes were estimated using Landsat TM and ETM+ imagery. With these satellite data and the BEPS process model (Boreal Ecosystem Productivity Simulator), the impacts of urbanization on regional net primary productivity (NPP) and annual net primary production were assessed for 1991 and 2002. Landsat-based land cover maps in 1991 and 2002 showed that urban development encroached large areas of cropland and forest. Expansion of residential areas and reduction of vegetated areas were the major forms of land transformation in Jiangyin County during this period. Mean NPP of the total area decreased from 818 to 699 gCm(-2)yr(-1) during the period of 1991 to 2002. NPP of cropland was only reduced by 2.7% while forest NPP was reduced by 9.3%. Regional annual primary production decreased from 808 GgC in 1991 to 691 GgC in 2002, a reduction of 14.5%. Land cover changes reduced regional NPP directly, and the increasing intensity and frequency of human-induced disturbance in the urbanized areas could be the main reason for the decrease in forest NPP.     

The Impact of Map and Data Resolution on the Determination of the Agricultural Utilisation of Organic Soils in Germany

Due to its nature, agricultural land use depends on local site characteristics such as production potential, costs and external effects. To assess the relevance of the modifying areal unit problem (MAUP), we investigated as to how a change in the data resolution regarding both soil and land use data influences the results obtained for different land use indicators. For the assessment we use the example of the greenhouse gas (GHG) emissions from agriculturally used organic soils (mainly fens and bogs). Although less than 5 % of the German agricultural area in use is located on organic soils, the drainage of these areas to enable their agricultural utilization causes roughly 37 % of the GHG emissions of the German agricultural sector. The abandonment of the cultivation and rewetting of organic soils would be an effective policy to reduce national GHG emissions. To assess the abatement costs, it is essential to know which commodities, and at what quantities, are actually produced on this land. Furthermore, in order to limit windfall profits, information on the differences of the profitability among farms are needed. However, high-resolution data regarding land use and soil characteristics are often not available, and their generation is costly or the access is strictly limited because of legal constraints. Therefore, in this paper, we analyse how indicators for land use on organic soils respond to changes in the spatial aggregation of the data. In Germany, organic soils are predominantly used for forage cropping. Marked differences between the various regions of Germany are apparent with respect to the dynamics and the intensity of land use. Data resolution mainly impairs the derived extent of agriculturally used peatland and the observed intensity gradient, while its impact on the average value for the investigated set of land-use indicators is generally minor.     

Characteristics of Urban Natural Areas Influencing Winter Bird Use in Southern Ontario,Canada

Characteristics of urban natural areas and surrounding landscapes were identified that best explain winter bird use for 28 urban natural areas in southern Ontario, Canada. The research confirms for winter birds the importance of area (size) and natural vegetation, rather than managed, horticultural parkland, within urban natural areas as well as percent urban land use and natural habitat in surrounding landscapes. Alien bird density and percent ground feeding species increased with percent surrounding urban land use. Higher percent forest cover was associated with higher percentages of forest, bark feeding, small (<20 g) and insectivorous species. Natural area size (ha) was related to higher species richness, lower evenness and higher percentages of insectivorous, forest interior, area-sensitive, upper canopy, bark feeding, and non-resident species. Higher number of habitat types within natural areas and percent natural habitat in surrounding landscapes were also associated with higher species richness. Common, resident bird species dominated small areas (<6.5 ha), while less common non-residents increased with area, indicative of a nested distribution. Areas at least 6.5 ha and more generally >20 ha start to support some area-sensitive species. Areas similar to rural forests had >25% insectivores, >25% forest interior species, >25% small species, and <5% alien species. Indicator species separated urban natural areas from rural habitats and ordination placed urban natural areas along a gradient between urban development and undisturbed, rural forests. More attention is needed on issues of winter bird conservation in urban landscapes.     

Assessing Watershed-Scale, Long-Term Hydrologic Impacts of Land-Use Change Using a GIS-NPS Model

Land-use change, dominated by an increase in urban/impervious areas, has a significant impact on water resources. This includes impacts on nonpoint source (NPS) pollution, which is the leading cause of degraded water quality in the United States. Traditional hydrologic models focus on estimating peak discharges and NPS pollution from high-magnitude, episodic storms and successfully address short-term, local-scale surface water management issues. However, runoff from small, low-frequency storms dominates long-term hydrologic impacts, and existing hydrologic models are usually of limited use in assessing the long-term impacts of land-use change. A long-term hydrologic impact assessment (L-THIA) model has been developed using the curve number (CN) method. Long-term climatic records are used in combination with soils and land-use information to calculate average annual runoff and NPS pollution at a watershed scale. The model is linked to a geographic information system (GIS) for convenient generation and management of model input and output data, and advanced visualization of model results. The L-THIA/NPS GIS model was applied to the Little Eagle Creek (LEC) watershed near Indianapolis, Indiana, USA. Historical land-use scenarios for 1973, 1984, and 1991 were analyzed to track land-use change in the watershed and to assess impacts on annual average runoff and NPS pollution from the watershed and its five subbasins. For the entire watershed between 1973 and 1991, an 18% increase in urban or impervious areas resulted in an estimated 80% increase in annual average runoff volume and estimated increases of more than 50% in annual average loads for lead, copper, and zinc. Estimated nutrient (nitrogen and phosphorus) loads decreased by 15% mainly because of loss of agricultural areas. The L-THIA/NPS GIS model is a powerful tool for identifying environmentally sensitive areas in terms of NPS pollution potential and for evaluating alternative land use scenarios for NPS pollution management.     

The impact of land use and land cover changes on land surface temperature in a karst area of China

Satellite images have been used extensively to study temporal changes in land use and land cover (LULC) in China. However, few studies have been conducted in the karst areas despite the large area and population involved and the fragile ecosystem. In this study, LULC changes were examined in part of Guizhou Province of southern China from 1991 to 2001 based on Landsat Thematic Mapper (TM) images of November 7, 1991, December 5, 1994, and December 19, 2001. Land surface temperature (LST) and normalized difference vegetation index (NDVI) were computed based on LULC types. The results show that agricultural land decreased, while urban areas expanded dramatically, and forest land increased slightly. Barren land increased from 1991 to 1994, and then decreased from 1994 to 2001. These changes in LULC widened the temperature difference between the urban and the rural areas. The change in LST was mainly associated with changes in construction materials in the urban area and in vegetation abundance both in the urban and rural areas. Vegetation had a dual function in the temperatures of different LULC types. While it could ease the warming trend in the urban or built-up areas, it helped to keep other lands warmer in the cold weather. The study also reveals that due to the government's efforts on reforestation, rural ecosystems in some of the study area were being restored. The time required for the karst ecosystem to recover was shorter than previously thought.     

Evaluation of the National Land Cover Database for Hydrologic Applications in Urban and Suburban Baltimore,Maryland1

Smith, Monica Lipscomb, Weiqi Zhou, Mary Cadenasso, Morgan Grove, and Lawrence E. Band, 2010. Evaluation of the National Land Cover Database for Hydrologic Applications in Urban and Suburban Baltimore, Maryland. Journal of the American Water Resources Association (JAWRA) 46(2):429-442. DOI: 10.1111/j.1752-1688.2009.00412.x Abstract: We compared the National Land Cover Database (NLCD) 2001 land cover, impervious, and canopy data products to land cover data derived from 0.6-m resolution three-band digital imagery and ancillary data. We conducted this comparison at the 1 km², 9 km², and gauged watershed scales within the Baltimore Ecosystem Study to determine the usefulness and limitations of the NLCD in heterogeneous urban to exurban environments for the determination of land-cover information for hydrological applications. Although the NLCD canopy and impervious data are significantly correlated with the high-resolution land-cover dataset, both layers exhibit bias at <10 and >70% cover. The ratio of total impervious area and connected impervious area differs along the range of percent imperviousness – at low percent imperviousness, the NLCD is a better predictor of pavement alone, whereas at higher percent imperviousness, buildings and pavement together more resemble NLCD impervious estimates. The land-cover composition and range for each NLCD urban land category (developed open space, low-intensity, medium-intensity, and high-intensity developed) is more variable in areas of low-intensity development. Fine-vegetation land-cover/lawn area is incorporated in a large number of land use categories with no ability to extract this land cover from the NLCD. These findings reveal that the NLCD may yield important biases in urban, suburban, and exurban hydrologic analyses where land cover is characterized by fine-scale spatial heterogeneity.     

SIMULATING HYDROLOGIC AND WATER QUALITY IMPACTS IN AN URBANIZING WATERSHED1

ABSTRACT: The Hydrologic Simulation Program‐Fortran (HSPF) was calibrated and used to assess the future effects of various land development scenarios on water quality in the Polecat Creek watershed in Caroline County, Virginia. Model parameters related to hydrology and water quality were calibrated and validated using observed stream flow and water quality data collected at the watershed outlet and the outlets of two sub water sheds. Using the county's Comprehensive Plan, land use scenarios were developed by taking into account the trends and spatial distributions of future development. The simulation results for the various land use scenarios indicate that runoff volume and peak rate increased as urban areas increased. Urbanization also increased sediment loads mainly due to increases in channel erosion. Constituent loads of total Kjeldal nitrogen, orthophosphorus, and total phosphorous for Polecat Creek watershed slightly decreased under future development scenarios. These reductions are due to increases in urban areas that typically contribute smaller quantities of nitrogen and phosphorous, as compared to agricultural areas. However, nitrate loads increased for the future land use scenarios, as compared to the existing land use. The increases in nitrate loads may result from increases in residential land and associated fertilizer use and concurrent decreases in forested land. The procedures used in this paper could assist local, state, and regional policy makers in developing land management strategies that minimize environmental impacts while allowing for future development.     

城市内区域土地集约利用综合评价探讨——以唐山市乐亭县为例

城市内区域土地集约利用评价能够反映各区域土地集约利用程度,为完善城市总体规划提供依据。以唐山市乐亭县为例,建立一套切实可行的城市内区域土地集约利用评价指标体系,运用多因素综合评价法进行评价。评价结果表明,新区高于旧区、中心区高于城市边缘区的集约特性,区域问土地集约利用程度差异巨大,城市土地利用的重点应放在已建成区土地的挖潜改造上,走内涵集约式发展道路,最大程度地提高城市土地利用效益。     

The Impact of Future Land Use Scenarios on Runoff Volumes in the Muskegon River Watershed

In this article we compared the response of surface water runoff to a storm event for different rates of urbanization, reforestation and riparian buffer setbacks across forty subwatersheds of the Muskegon River Watershed located in Michigan, USA. We also made these comparisons for several forecasted and one historical land use scenarios (over 140 years). Future land use scenarios to 2040 for forest regrowth, urbanization rates and stream setbacks were developed using the Land Transformation Model (LTM). Historical land use information, from 1900 at 5-year time step intervals, was created using a Backcast land use change model configured using artificial neural network and driven by agriculture and housing census information. We show that (1) controlling the rate of development is the most effective policy option to reduce runoff; (2) establishing setbacks along the mainstem are not as effective as controlling urban growth; (3) reforestation can abate some of the runoff effects from urban growth but not all; (4) land use patterns of the 1970s produced the least amount of runoff in most cases in the Muskegon River Watershed when compared to land use maps from 1900 to 2040; and, (5) future land use patterns here not always lead to increased (worse) runoff than the past. We found that while ten of the subwatersheds contained futures that were worse than any past land use configuration, twenty-five (62.5%) of the subwatersheds produced the greatest amount of runoff in 1900, shortly after the entire watershed was clear-cut. One third (14/40) of the subwatersheds contained the minimum amount of runoff in the 1960s and 1970s, a period when forest amounts were greatest and urban amounts relatively small.     

Forecasts of habitat loss and fragmentation due to urban growth are sensitive to source of input data

The conversion of natural habitat to urban settlements is a primary driver of biodiversity loss, and species' persistence is threatened by the extent, location, and spatial pattern of development. Urban growth models are widely used to anticipate future development and to inform conservation management, but the source of spatial input to these models may contribute to uncertainty in their predictions. We compared two sources of historic urban maps, used as input for model calibration, to determine how differences in definition and scale of urban extent affect the resulting spatial predictions from a widely used urban growth model for San Diego County, CA under three conservation scenarios. The results showed that rate, extent, and spatial pattern of predicted urban development, and associated habitat loss, may vary substantially depending on the source of input data, regardless of how much land is excluded from development. Although the datasets we compared both represented urban land, different types of land use/land cover included in the definition of urban land and different minimum mapping units contributed to the discrepancies. Varying temporal resolution of the input datasets also contributed to differences in projected rates of development. Differential predicted impacts to vegetation types illustrate how the choice of spatial input data may lead to different conclusions relative to conservation. Although the study cannot reveal whether one dataset is better than another, modelers should carefully consider that geographical reality can be represented differently, and should carefully choose the definition and scale of their data to fit their research objectives.     

ANALYSIS AND PREDICTION OF SURFACE RUNOFF IN AN URBANIZING WATERSHED USING SATELLITE IMAGERY1

ABSTRACT: This paper demonstrates how satellite image data [e.g., from Landsat 5 Thematic Mapper (TM)], in conjunction with an urban growth model and simple runoff calculations, can be used to estimate future surface runoff and, by implication, water quality within a watershed. To illustrate the method, predictions of land use change and surface runoff are shown for Spring Creek Watershed, a medium sized urbanizing watershed in Central Pennsylvania. Land cover classifications for this watershed were created from images for summertime 1986 and 1996 and subsequently used as input to the Clarke urban growth model, called SLEUTH, to predict land use changes to the year 2025. Simulations with this model show a progressive growth in the percentage of urban pixels and in impervious surface area in the watershed but also an increase in woodland, primarily in previously clear‐cut areas. Given that woodland area will continue to increase in area, surface runoff into Spring Creek is predicted to remain only slightly above present level. However, should the woodland amount fail to increase, surface runoff is then predicted to increase more significantly during the next 25 years. Finally, the concept of urban sprawl is addressed within the context of predicted increases in urbanization by relating the implied increase in impervious surface area to population density within the watershed.     

Analyses of Urban Drainage Network Structure and its Impact on Hydrologic Response1

Meierdiercks, Katherine L., James A. Smith, Mary Lynn Baeck, and Andrew J. Miller, 2010. Analyses of Urban Drainage Network Structure and Its Impact on Hydrologic Response. Journal of the American Water Resources Association (JAWRA) 1-12. DOI: 10.1111/j.1752-1688.2010.00465.x Abstract: Urban flood studies have linked the severity of flooding to the percent imperviousness or land use classifications of a watershed, but relatively little attention has been given to the impact of urban drainage networks on hydrologic response. The drainage network, which can include storm pipes, surface channels, street gutters, and stormwater management ponds, is examined in the Dead Run watershed (14.3 km²). Comprehensive digital representations of the urban drainage network in Dead Run were developed and provide a key observational resource for analyses of urban drainage networks and their impact on hydrologic response. Analyses in this study focus on three headwater subbasins with drainage areas ranging from 1.3 to 1.9 km² and that exhibit striking contrasts in their patterns and history of development. It is shown that the drainage networks of the three subbasins, like natural river networks, exhibit characteristic structures and that these features play critical roles in determining urban hydrologic response. Hydrologic modeling analyses utilize the Environmental Protection Agency’s Stormwater Management Model (SWMM), which provides a flexible platform for examining the impacts of drainage network structure on hydrologic response. Results of SWMM modeling analyses suggest that drainage density and presence of stormwater ponds impact peak discharge more significantly in the Dead Run subbasins than the percent impervious or land use type of the subbasins.     

Arsenic distribution in Florida urban soils: comparison between Gainesville and Miami

Arsenic contamination is of concern due to its effect as a carcinogen. Understanding the distribution of arsenic in urban soils is important for establishing baseline concentrations from which anthropogenic effects can be measured. The soil cleanup target level (SCTL) for arsenic in Florida (0.8 and 3.7 mg kg(-1) in residential and commercial areas, respectively) is lower than in most states and is near the arsenic background concentrations in Florida soils. The objective of this study was to characterize the distribution of arsenic in the soils of two Florida cities, Gainesville and Miami. More than 200 soil samples were collected from three land-use classes in each city (residential, commercial, and public land), digested with USEPA Method 3051a, and analyzed with graphite furnace atomic absorption spectrophotometry. Arsenic concentrations varied greatly in Gainesville, ranging from 0.21 to approximately 660 mg kg(-1) with a geometric mean (GM) of 0.40 mg kg(-1) (after discarding outliers), which was significantly lower than the GM of 2.81 mg kg(-1) in Miami, although Miami samples ranged only from 0.32 to approximately 110 mg kg(-1). Arsenic concentrations in 29 and 4% of the Gainesville soil samples and 95 and 33% of the Miami samples exceeded the Florida residential and commercial SCTL, respectively. This study is the first to provide information on arsenic distribution in urban soils of Florida, and the data are useful for assessing arsenic contamination and determining the need for remediation.     

USING LANDSAT DATA TO CLASSIFY LAND USE FOR ASSESSING THE BASINWIDE RUNOFF INDEX1

Remote sensing data in the form of Landsat computer compatible tapes (CCT) was used to determine land use and land cover as an aid in hydrologic studies that were used to estimated a basinwide runoff index. With the use of the General Electric Image 100 multispectral image processing system in conjunction with the Earth Resources Laboratory Application Software (ELAS), CCT's on February 9, 1976, were analyzed by spectral differences to determine unique land use conditions within the Econlockhatchee (Econ) River Basin, Florida. The result showed that the Landsat data can be successfully used to monitor the USGS land use Level 1. An advantage of using the Landsat data for land use classification is that new data are periodically available for updating the land use information. The Soil Conservation Service curve number was used to establish a basinwide runoff index which includes a prime variable of land use changes with the time. The basinwide runoff index in 1972 (with USGS 1972 Land Use maps) was similar to the one in 1976 (with Landsat data dated February 9, 1976). This implies that the runoff from the entire Econ Basin was not noticeably changed during the period of 1972 and 1976.     

Visual interpretation of satellite imagery for monitoring floods in Bangladesh

This article presents the methods and results of visual interpretation of NOAA AVHRR imagery for mapping flood areas in Bangladesh. Color composites of channels 1 and 2 for 18 August 1987 and 10, 15, and 24 September 1988 were interpreted visually for delineating flood boundaries. On such imagery flood areas appear in dark tones and are separated from land (light tones) owing to the absorption of near infrared by water and its reflectance by land and non-waterbodies. Visual interpretation was aided by the use of ground information, such as physiographic and river maps, previous flood maps, newspaper reports, and other published and unpublished documents on the 1987 and 1988 floods. Interpreted flood areas on selected images ranged from 31% to 43% of the total area of Bangladesh. Visual interpretation overestimated flood areas by 5%–10%, compared to the digitally analyzed data. The main advantage of visual analysis lies in the cost effectiveness of AVHRR photographic products, which make them more accessible than the digital image analysis of computer-compatible tapes.     

Impact of Urban Sprawl on Water Quality in Eastern Massachusetts,USA

A study of water quality, land use, and population variations over the past three decades was conducted in eastern Massachusetts to examine the impact of urban sprawl on water quality using geographic information system and statistical analyses. Since 1970, eastern Massachusetts has experienced pronounced urban sprawl, which has a substantial impact on water quality. High spatial correlations are found between water quality indicators (especially specific conductance, dissolved ions, including Ca, Mg, Na, and Cl, and dissolved solid) and urban sprawl indicators. Urbanized watersheds with high population density, high percentage of developed land use, and low per capita developed land use tended to have high concentrations of water pollutants. The impact of urban sprawl also shows clear spatial difference between suburban areas and central cities: The central cities experienced lower increases over time in specific conductance concentration, compared to suburban and rural areas. The impact of urban sprawl on water quality is attributed to the combined effects of population and land-use change. Per capita developed land use is a very important indicator for studying the impact of urban sprawl and improving land use and watershed management, because inclusion of this indicator can better explain the temporal and spatial variations of more water quality parameters than using individual land use or/and population density.     

Landform Classification for Land Use Planning in Developed Areas: An Example in Segovia Province (Central Spain)

Landform-based physiographic maps, also called land systems inventories, have been widely and successfully used in undeveloped/rural areas in several locations, such as Australia, the western United States, Canada, and the British ex-colonies. This paper presents a case study of their application in a developed semi-urban/suburban area (Segovia, Spain) for land use planning purposes. The paper focuses in the information transfer process, showing how land use decision-makers, such as governments, planners, town managers, etc., can use the information developed from these maps to assist them. The paper also addresses several issues important to the development and use of this information, such as the goals of modern physiography, the types of landform-based mapping products, the problem of data management in developed areas, and the distinctions among data, interpretations, and decisions.     

A NONTRADITIONAL METHODOLOGY FOR FLOOD STAGE-DAMAGE CALCULATIONS1

ABSTRACT: This paper presents a new methodology to calculate economic losses from hypothetical, extreme flood events, such as the Probable Maximum Flood. The methodology uses economic data compiled from already-available secondary sources, such as U.S. Census data on magnetic tapes, utilizing microcomputer and other electronic media. Estimates of land elevations are obtained from topographic maps, and flood elevations axe estimated using, for example, a dam breach and flood routing (DAMBRK) model (Fread, 1984). The calculations are performed at a disaggregate spatial scale, by various land use and industrial classification categories. The basic areal units are city blocks (for urbanized areas), enumeration districts, and Census tracts. Depth-damage functions, which provide an estimate of damages as a proportion of the existing value of the structure, are estimated statistically. Computer software (called DAMAGE) is used to combine the economic, flood elevation, and depth-damage information to compute economic losses for different possible flood stages and for different inflow events. Two case studies are presented as illustrations of the method.     

城市土地集约利用潜力的分类与分阶

研究城市土地集约利用潜力的类别与阶次细分,为城市土地集约利用挖潜的具体步骤提供宏观导向。研究结果表明,按照潜力的属性,应将城市土地集约利用潜力分为面积潜力、经济潜力;在时间尺度上,按挖潜程度的难易,城市土地集约利用潜力可分为一阶、二阶、三阶、四阶。城市土地集约利用潜力评价应将潜力测算的结果进行分类分阶,这样才能使城市土地集约挖潜工作落到实处,可操作性更强。