Literature Review

Home » Research » Environmental Studies » LULC Change in Carroll County » Literature Review

Other LULC Change in Carroll County areas:

Loading, please wait...

Chapter 3 - Literature Review

Land Use & Land Cover (return to top)

Every parcel of land on the Earth’s surface is unique in the cover it possesses. Land use and land cover are distinct yet closely linked characteristics of the Earth’s surface. Land use is the manner in which human beings employ the land and its resources. Examples of land use include agriculture, urban development, grazing, logging, and mining. In contrast, land cover describes the physical state of the land surface. Land cover categories include cropland, forests, wetlands, pasture, roads, and urban areas. The term land cover originally referred to the kind and state of vegetation, such as forest or grass cover, but it has broadened in subsequent usage to include human structures such as buildings or pavement and other aspects of the natural environment, such as soil type, biodiversity, and surface and groundwater (Meyer, 1995).

Land Use & Land Cover Change (return to top)

Land use affects land cover and changes in land cover affect land use. A change in either, however, is not necessarily the product of the other. Changes in land cover by land use do not necessarily imply a degradation of the land. However, many shifting land use patterns, driven by a variety of social causes, result in land cover changes that affect biodiversity, water and radiation budgets, trace gas emissions and other processes that, cumulatively, affect global climate and biosphere (Riebsame, Meyer, and Turner, 1994).

Land cover can be altered by forces other than anthropogenic. Natural events such as weather, flooding, fire, climate fluctuations, and ecosystem dynamics may also initiate modifications upon land cover. Globally, land cover today is altered principally by direct human use: by agriculture and livestock raising, forest harvesting and management, and urban and suburban construction and development. There are also incidental impacts on land cover from other human activities such as forests and lakes damaged by acid rain from fossil fuel combustion and crops near cities damaged by tropospheric ozone resulting from automobile exhaust (Meyer, 1995).

Contemporary global change consists of two broad types, systemic and cumulative. Systemic change operates directly on the bio-chemical flows that sustain the biosphere and, depending on its magnitude, can lead to global change, just as fossil fuel consumption increases the concentration of atmospheric carbon dioxide. Systemic change is largely associated with, but not limited to, the Industrial Age and thus has grown especially important over the more recent past (Turner and Butzer, 1992).

Cumulative change has been the most common type of human induced environmental change since antiquity. Cumulative changes are geographically limited, but if repeated sufficiently, become global in magnitude. Changes in landscape, cropland, grasslands, wetlands, or human settlements are examples of cumulative change. Some cumulative changes reached continental, even global, proportions long before the 20th Century, including deforestation and the modification of grasslands (Turner and Butzer, 1992).

Changes in land cover driven by land use can be categorized into two types: modification and conversion. Modification is a change of condition within a cover type; for example, unmanaged forest modified to a forest managed by selective cutting. Significant modifications of land cover can occur within these patterns of land cover change. Conversion is a change from one cover type to another, such as deforestation to create cropland or pasture. Conversion land cover changes such as deforestation have been the focus of many global change research agendas (Riebsame, Meyer and Turner, 1994).

The loss of rainforests throughout the tropical regions of the world as a result of deforestation for timber resources and conversion to agricultural lands has become a topic of global attention with the aid of widespread media coverage. Research specialists such as Skole & Tucker (1993), Skole et al (1994), and Kummer & Turner (1994) perform extensive studies in an attempt to bring further attention to this situation by focusing on the social implications and the environmental degradation associated with tropical deforestation in the Amazon of South America and in Southeast Asia. Yet, with all the research, awareness, and attention of the world, this potentially devastating phenomenon continues. It is an unfortunate, but fact of life that deforestation occurs on numerous expanses and at varying scales around the globe. Our society’s focus on the plight of the tropical rainforests appears to have overlooked the shifting patterns of land use and land cover occurring in our own forests. This research focuses on the conversion of forest to pasture in a localized, rural setting in hopes that awareness of such occurrences may also be further publicized.

Multispectral Classification & Mapping of the Landscape (return to top)

The process of multispectral mapping of the landscape consists of drawing boundaries around geographically located classes that are homogeneous, or acceptably heterogeneous, and the description of those classes and their attributes and relations in a consistent and logical manner (Robinove, 1981). The Earth’s landscape, considered in the broadest sense, has an extremely large number of attributes that may be used for classification and description, depending upon the purpose of the classification and the needs of the classifier.

The use of multispectral reflectance data for mapping land use and land cover has become an integral component of contemporary land use studies. Useful information such as vegetation maps and detailed soil maps have not always been readily available in a digital format. In many areas there is still great demand for localized digital data. Therefore, land use and land cover mapping procedures often rely heavily on the differences of spectral characteristics of the landscape for separation into meaningful land use and land cover classes. Multispectral reflectance data, or remotely sensed imagery, from satellite sensors serve as surrogate data representative of landscape features or attributes. The utilization of remotely sensed data enables surrogate mapping due to the impracticality of direct measurement of the landscape (Robinove, 1981).

Remote Sensing & Innovative Mapping Technologies (return to top)

Remote Sensing is the science and art of obtaining information about an object, area, or phenomenon through the analysis of data acquired by a device that is not in contact with object, area, or phenomenon under investigation (Lillesand and Kiefer, 1987). Within the scope of this study, the focus of remote sensing is the measurement of emitted or reflected electromagnetic radiation, or spectral characteristics, from a target object by a multispectral satellite sensor. A multispectral sensor is characterized as a passive sensor. Passive sensors record energy that is naturally reflected or emitted from a target. In contrast, active sensors supply their own source of energy, directing it at the target in order to measure the returned energy.

A multispectral sensor acquires multiple images of the same target object at different wavelengths (bands). Each band measures unique spectral characteristics about the target. A spectral band is a data set collected by the sensor with information from discrete portions of the electromagnetic spectrum. The electromagnetic (EM) spectrum is a range of electromagnetic radiation ranging from cosmic waves to radio waves. Multispectral sensors focus on ranges on the EM spectrum where radiation penetrates the air with little or no loss by absorption of the target. Remote sensors on space platforms are programmed to operate in these windows and make measurements using detectors tuned to these specific wavelength frequencies which pass through the atmosphere. Spectral reflectance characteristics of common earth surface materials are located within the visible and near to mid-infrared range (Richards, 1986).

In most contemporary land use studies which employ remote sensing imagery from multispectral sensors, the foremost task is the observation of spectral characteristics of measured electromagnetic radiation from a target or landscape. Analysts develop signatures based upon the detected energy’s measurement and position in the electromagnetic spectrum. A signature is a set of statistics that defines the spectral characteristic of a target phenomenon or training-sites. Image analysts determine the measurement of signature separability by determining quantitatively the relation between class signatures. Signatures are refined by improved ground-truth and accuracy assessment analysis. By utilizing the developed signatures in mulitspectral classification and thematic mapping, the analyst generates new data for analysis (ERDAS, 1999).

Today, remote sensing image data of the Earth’s surface acquired by spacecraft platforms is readily available in a digital format. Digital remote sensing systems convert electromagnetic energy (color, light, heat, etc…) to a digital form. Spatially, the data is composed of discrete picture elements, or pixels, and radiometrically it is quantised into discrete brightness levels (ERDAS, 1999). The great advantage of having data available digitally is that it can be processed by computer either for machine assisted information extraction or for the enhancement by an image interpreter.

Resolution is an important term commonly used to describe remotely sensed imagery. However, there are four distinct types of resolution that must be considered. These four types of resolution are spatial, spectral, radiometric, and temporal. These resolution characteristics help to describe the functionality of both remote sensing sensors and remotely sensed data. The ERDAS, Earth Resources Data Analysis System, Field Guide (1999) further describes each type of resolution.

Spatial resolution is the minimum size of terrain features that can be distinguished from the background in an image, or the ability to differentiate between two closely spaced features in an image. It is also defined by the area on the ground that a pixel represents in a digital image file. Large scale in remote sensing refers to imagery in which each pixel represents a small area on the ground. Small scale refers to imagery in which each pixel represents a large area on the ground.

Spectral resolution refers to the number and dimension of specific wavelength intervals in the electromagnetic spectrum to which a sensor or sensor band is sensitive or can record. Wide intervals in the electromagnetic spectrum are referred to as coarse spectral resolution, and narrow intervals are referred to as fine spectral resolution.

Radiometric resolution refers to the dynamic range, or number of possible data file values in each band. This is referred to by the number of bits into which the recorded energy is divided. The total intensity of the energy, from 0 to the maximum amount, the sensor measures is broken down, for example, into 256 brightness values for 8-bit data. The data file values range from 0, for no energy return, to 255, for maximum return, for each pixel.

Temporal resolution is a measure of how often a given sensor system obtains imagery of a particular area, or how often an area can be revisited. The temporal resolution of satellites are on a fixed schedule. The fixed schedule of satellites allows for more repetitive views. This revisit capability makes it possible to use several passes, perhaps covering two or three seasons or multiple years, for interpretation. In addition, new satellite technology is incorporating pointable or directional sensors allowing for even quicker revisit capabilities. Temporal resolution is an important factor to consider in change detection studies. Landsat 5 can view the same area of the globe every 16 days (Wilkie & Finn, 1996).

Remote sensing has become an important tool applicable to developing and understanding the global, physical processes affecting the earth. As current trends continue, additional and higher resolution satellites will become available providing the means to produce more accurate land use and land cover maps characterized by finer levels of detail. Radiometric resolution refers to the dynamic range, or number of possible data file values in each band. This is referred to by the number of bits into which the recorded energy is divided. The total intensity of the energy is broken down into 256 brightness values for 8-bit data. The data file values range from 0 to 255 for each pixel, with 0 being no return and 255 maximum return. The sensor measures the electromagnetic radiation in this range (ERDAS, 1994). Landsat TM imagery is classified as 8-bit data.

Landsat Thematic Mapper Sensor & Multispectral Imagery (return to top)

Lauer et al (1997) give a brief history of the Landsat program and it’s renowned success. The United States has pioneered land remote sensing from space and has been the leader in the development of earth observing technology over the past thirty years. The evolution of the Landsat program has been a fundamental genesis to an international endeavor to better measure and monitor the Earth and its precious resources. Despite early military/intelligence programs in space during the 1950’s and 60’s, the scientific and industrial communities in the U.S. became aware of the potential of earth observing vehicles in space. The National Aeronautics and Space Administration, NASA, in cooperation with other federal agencies, successfully launched on July 23, 1972, the first Earth Resources Technology Satellite (ERTS-1), which was later renamed Landsat 1. Landsat 1 was a Nimbus-type platform which carried a sensor package and data-relay equipment. ERTS-2 was launched on January 22, 1975, and was also renamed Landsat 2. Additional Landsats were launched in 1978, 1982, and 1984 (Landsats 3, 4, and 5 respectively). Each successive satellite system has had improved sensor and communication capabilities.

The Landsat program has had an enormous impact in numerous application arenas. In addition to the inauguration of global research, the Landsat program has also provided researchers with real-world data and access to greatly enhanced spatial and analytical tools. The premise of the Landsat program is that the Earth’s features and landscapes can be discriminated, identified, categorized, and mapped on the basis of their spectral reflectances and emissions. The Landsat program has continued to provide a flow of high quality, well-calibrated, synoptic imagery of the earth and has opened new insights into geologic, agricultural, and land use surveys, in addition to new paths in resource exploration. An understanding of the Earth and its terrestrial ecosystems, as well as its land processes, has been advanced remarkably by the Landsat program and will continue to do so into the future with the launch of the program’s most recent satellite platform, Landsat 7 on April, 15, 1999.

Data incorporated in this project was obtained from the Thematic Mapper sensor onboard both Landsat 5 and Landsat 7 satellite sensors. Landsat 5 has been in service since March 1,1984 and maintains an orbit altitude of 705 kilometers above the surface of the earth. The principal observing instruments on Landsat 5 and Landsat 7 is the Thematic Mapper (TM) and the Multispectral Scanner (MSS). The Thematic Mapper has considerably greater spatial, spectral, and radiometric resolution than the multispectral scanner. The Thematic Mapper has seven spectral bands with an instantaneous field of view at a nadir of 30 meters, spatial resolution, except for the thermal band with a field of view of 120 meters which was not incorporated in this research. In addition, the Thematic Mapper has 256 quantized levels characterized by 8 bit data.

These sensors provide a 16-day, 233 orbit repeat cycle with image sidelap that varies from 14 percent at the equator to nearly 84 percent at 81 degrees North or South latitude. (Figure 3.1) Previous Landsat sensors, 1 – 3, required 18-day and 251 overlapping orbits. The sensors on these satellites obtain images with a swath width of 185 kilometers (Lauer et al, 1997).

The Landsat World-Wide-Reference system catalogues the Earth’s landmasses into 57,784 scenes, each 115 miles (185 kilometers) wide by 106 miles (170 kilometers) long (USGS, 1999). An example of the approximate coverage extent of a Landsat TM scene can be seen in figure 3.2. It requires 10 separate Landsat Scenes mosaiced or tiled together in order to cover the entire state of Arkansas. The study area considered for this project, Carroll County shown in red, is entirely contained in the Landsat scene Path/Row 25/35. (Figure 3.2)

Figure 3.1

Figure 3.2

Landsat Thematic Mapper is a passive, multispectral sensor that measures the strength of emitted or reflected electromagnetic radiation. The strength of the received signal is quantified or indexed using Digital Numbers (DN’s). An array of Digital Numbers are displayed as picture elements (Pixels), and when combined, multiple pixels form digital images of the landscape (ERDAS, 1999). The ERDAS Field Guide (1999) also gives additional insight describing the spectral resolution of Landsat Thematic Mapper. The sensors of the Thematic Mapper record electromagnetic radiation in seven bands. Bands 1, 2, and 3 are in the visible portion of the spectrum. Bands 4, 5, and 7 are in the reflective-infrared portion of the spectrum. Band 6 is in the thermal portion of the spectrum. The following list describes each of the seven TM bands:

Band 1 – Visible Blue, 0.45 – 0.52 um Useful for mapping coastal water areas, differentiating between soil and vegetation, forest type mapping, and detecting cultural features.

Band 2 – Visible Green, 0.52 – 0.60 um Corresponds to the green reflectance of healthy vegetation. Also useful for cultural feature identification.

Band 3 – Visible Red, 0.63 – 0.69 um Useful for discriminating between many plant species. It is also useful for determining soil boundary and geological boundary delineations as well as cultural features.

Band 4 – Reflective - infrared, 0.76 – 0.90 um This band is especially responsive the amount of vegetation biomass present in a scene. It is useful for crop identification and emphasizes soil/crop and land/water contacts.

Band 5 – Mid - infrared, 1.55 – 1.74 um This band is sensitive to the amount of water in plants, which is useful in crop drought studies and in plant health analyses. This is also one of the few bands that can be used to discriminate between clouds, snow, and ice.

Band 6 – Thermal - infrared, 10.40 – 12.50 um This band is useful for vegetation and crop stress detection, heat intensity, insecticide applications, and for locating thermal pollution. It can also be used to locate geothermal activity.

Band 7 – Mid - infrared, 2.08 – 2.35 um This band is important for the discrimination of geologic rock type and soil boundaries, as well as soil and vegetation moisture content.

Different combinations of the TM bands can be displayed to create different composite effects. The following combinations are commonly used to display images:

* Bands 3, 2, and 1 create a true color composite. True color means that objects look as though they would to the naked eye, similar to a photograph.

* Bands 4, 3, and 2 create a false color composite. False color composites appear similar to an infrared photograph where objects do have the same colors or contrasts as they would naturally. For instance, in an infrared image, vegetation appears red, water appears navy or black.

* Bands 5, 4, and 2 create a pseudo color composite. (A thematic image is also a pseudo color image.) In pseudo color, the colors do not reflect the features in natural colors. For instance, roads may be red, water yellow, and vegetation blue.

With the adequate knowledge of band properties and the appropriate combination of Landsat TM bands, the extraction of numerous themes, land use and land cover classes, can be achieved for various mapping applications.

Land Use & Land Cover Classification Systems (return to top)

A primary component of mapping land use and land cover is adopting or developing a land cover classification system. Many current land use and land cover classification systems are designed specifically for use with remotely sensed data. Many of these classification systems often resemble or incorporate other classification systems in order to maintain cohesiveness and allow for data integration. A hierarchical framework is often implemented within a classification system. This type of framework allows the level of detail to vary for different project scopes and for the creation land use and land cover categories that are compatible with other classification systems(Foti et al, 1994).

Anderson et al (1967) developed a hierarchical land use and land cover classification system for utilization with remote sensor data which has been adopted by the U.S. Geological Survey for 1:250,000 and 1:100,000 scale land use and land cover mapping of the United States. The Anderson classification system and or Anderson derived land use and land cover classifications have been adopted in most contemporary land use and land cover research utilizing remotely sensed satellite data.

A hierarchical classification framework is composed of different levels of land use categories which are dependent upon the level of detail required by the project scope. Level I categories are often broad classification categories with examples such as urban, forest, agricultural land, and water classes. Level II categories offer more detail and are usually subdivisions of the level I categories. Examples of level II categories are coniferous forest, deciduous forest, and possibly mixed forest classes. Level III categories are often employed in local studies which incorporate species level land classes such as oak hickory forest or oak hickory pine forest (Anderson et al, 1976).

Within the scope of this study, level I and level II categories derived from the Anderson classification system compose the hierarchical land use and land cover classification of Carroll County and is further discussed in Chapter Four, Methodology.

Image Classification Techniques (return to top)

In contemporary land use and land cover mapping studies, land cover classes are often derived from satellite imagery by utilizing computer-assisted image classification techniques. Within the scope of the study, image classification is defined as the extraction of distinct classes or themes, land use and land cover classification categories, from satellite imagery. There are two primary methods of image classification utilized by image analysts, unsupervised and supervised classification.

Unsupervised image classification is a method in which the image interpreting software separates the pixels in an image based upon their reflectance values into classes or clusters with no direction from the analyst. Once this process is completed, the image analyst determines the land cover type for each class based on image interpretation, ground truth information, maps, field reports, etc… and assigns each class to a specified category by aggregation (ERDAS, 1999).

Supervised image classification is a method in which the analyst defines small areas, called training sites, on the image which are representative of each desired land cover category. The delineation of training areas representative of a cover type is most effective when an image analyst has knowledge of the geography of a region and experience with the spectral properties of the cover classes (Skidmore, 1989). The image analyst then trains the software to recognize spectral values or signatures associated with the training sites. After the signatures for each land cover category have been defined, the software then uses those signatures to classify the remaining pixels (ERDAS, 1999).

When classifying satellite imagery, single supervised or unsupervised classification techniques are often not enough to effectively classify an image. Automated classification accuracies can often be unacceptably low, < 80%, at the required level of categorical detail for many applications (Bolstad & Lillesand, 1992). Modifications of image classification techniques are most often required in order to assess for classification accuracy. Experimentation with proven or standardized classification techniques can produce accurate land cover classes as well as lead to the development of new classification procedures. Modifications of image classification techniques are often required in order to obtain an adequate classification accuracy. The classification methodology employed in this study is discussed further in Chapter Four, Methodology.

The accuracy of spatial data has been defined by the United States Geological Survey as: "The closeness of results of observations, computations, or estimates to the true values or the values accepted as being true" (USGS, 1990). Accuracy assessment or validation is an important step in the processing of remote sensing data. It determines the information value of the resulting data to a user. Productive utilization of geo-data is only possible if the quality of the data is known. Furthermore, integrated processing of different types of geo-data cannot be effective if the data quality is not known. During the last two decades, numerous studies have been published concerning accuracy assessment of land cover classifications (Congalton, 1996, Rosenfield & Fitzpatrick-Lins, 1986, and Foody, 1992).

The error matrix and kappa coefficient have become a standard means of assessment of image classification accuracy. This method of determining image classification accuracy resamples classified imagery against ground truth field samples often obtained with a Global Positioning System (GPS). Accuracy assessment is further discussed in Chapter 5, Results and Analysis.

Global Positioning Systems (GPS) (return to top)

Global Positioning Systems (GPS) provide the mapping community with powerful tools for acquiring accurate and current digital data. Combined with high resolution remote sensing and GIS for land use studies, GPS can provide high accuracy ground-truth data for training-site development.

The acquisition of ground truth data for this research project was made possible by the utilization of the Navstar, Navigation System with Timing and Ranging, Global Positioning System. This satellite navigation system, created in 1973 and operated by the U.S. Department of Defense, is an all-weather, 24-hour system that provides accurate three-dimensional position, velocity, and time. The Navstar GPS evolved from U.S. Air Force and Navy programs initiated in the mid-1960’s after recognizing the potential for mapping, charting, and geodesy (Heuerman & Senus, 1983).

Today, the Navstar GPS system is composed of a full constellation of 24 Block II satellites operating at an orbit of 12,600 miles above the earth. Navstar GPS features such as global reach, use of a common grid, immunity to saturation, and insensitivity to weather are critical elements for a reliable positioning system. GPS computations use the World Geodetic System of 1984. Since the system is passive, users only receive signals, the system cannot be saturated by too many users.

Essentially, the GPS satellites broadcast a continuously available time signal using an on-board atomic clock. Receivers use these time signals to calculate the distance to the satellite to establish an accurate position. However, the accuracy of GPS positions can vary substantially and the user must be aware of the factors that influence the precision of a GPS signal. The type of GPS service accessed, the type of GPS equipment and processing techniques utilized, and satellite geometry are some of the key factors affecting GPS precision (Bobbe, 1992). The utilization of GPS for the acquisition of ground truth field data for this project is discussed further in Chapter Four, Methodology.

Geographic Information Systems (GIS) (return to top)

Another recent development in the use of satellite data is to take advantage of increasing amounts of geographical data available in conjunction with geographic information systems to assist in interpretation. Geographical data describe objects from the real world in terms of (a) their position with respect to a known coordinate system,(b) their attributes that are unrelated to position (such as color, type, cost, pH, incidence of disease, etc.) and (c) their spatial interrelations with each other (topological relations), which describe how they are linked together or how one can travel between them (Burrough, 1986).

The concept of the geographic information system emerged during the 1960’s and 1970’s as new trends arose in the means in which maps were being produced and used for resource assessment, land evaluation, and planning. Essentially, this concept focuses on the ability to develop a powerful set of tools for collecting, storing, retrieving at will, transforming, and displaying spatial geographic data from the real world for specific analysis and inquiry. This set of tools constitutes a geographical information system or geographic information system. Geographic information systems are comprised of three main components: computer hardware, sets of application software modules, and a proper organization context (Burrough, 1986).

In addition, Robinove (1986) defines a geographic information system as a collection of computer programs in a given hardware environment which operate on a geographic database to analyze individual database elements or for synthesis of multiple database elements.

With the increasingly widespread, combined implementation of remote sensing and GIS technology, more natural resource professionals have been provided with efficient and accurate tools for mapping and maintaining management information on forests and other natural resources in regional areas. GIS technology is expanding, allowing for greater integration of remote sensing with digital cartography, thus providing the means to produce more accurate land use and land cover maps.

This research implements the power GIS to make effective use of digital terrain data and detailed soil maps to assess the wisdom and sustainability of current land use practices in Carroll County, Arkansas. Current land use practices are derived from a land use and land cover change map created from remotely sensed imagery. The utilization of GIS in this research is further discussed in Chapter Four, Methodology.

Land Use & Land Cover Change Detection (return to top)

An increasingly common application of remotely sensed data is for change detection. Change detection is the process of identifying differences in the state of an object or phenomenon by observing it at different times (Singh, 1989). Change detection is an important process in monitoring and managing natural resources and urban development because it provides quantitative analysis of the spatial distribution of the population of interest. Change detection is useful in such diverse applications as land use change analysis, monitoring shifting cultivation, assessment of deforestation, study of changes in vegetation phenology, seasonal changes in pasture production, damage assessment, crop stress detection, disaster monitoring, day/night analysis of thermal characteristics as well as other environmental changes (Singh, 1989).

Macleod and Congalton (1998) list four aspects of change detection which are important when monitoring natural resources:

1: Detecting that changes have occurred

2: Identifying the nature of the change

3: Measuring the areal extent of the change

4: Assessing the spatial pattern of the change

The basic premise in using remote sensing data for change detection is that changes in land cover result in changes in radiance values which can be remotely sensed. Techniques to perform change detection with satellite imagery have become numerous as a result of increasing versatility in manipulating digital data and increasing computing power.

A wide variety of digital change detection techniques have been developed over the last two decades. Singh (1989) and Coppin & Bauer (1996) both provide excellent and comprehensive summaries of methods and techniques of digital change detection. Coppin & Bauer (1996) summarize eleven different change detection algorithms that were found to be documented in the literature by 1995. These include:

1: monotemporal change delineation

2: delta or post-classification comparison

3: multidimensional temporal feature space analysis

4: composite analysis

5: image differencing

6: image ratioing

7: multitemporal linear data transformation

8: change vector analysis

9: image regression

10: multitemporal biomass index

11: background subtraction

The scientific literature reveals that digital change detection is a difficult task to perform accurately and unfortunately, many of the studies concerned with comparative evaluation of these applications have not supported their conclusions by quantitative analysis (Singh, 1989). All digital change detection is affected by spatial, spectral, temporal, and thematic constraints. The type of method implemented can profoundly affect the qualitative and quantitative estimates of the change. Even in the same environment, different approaches may yield different change maps. The selection of the appropriate method therefore takes on considerable significance. Not all detectable changes, however, are equally important to the resource manager. On the other hand, it is also probable that some changes of interest will not be captured very well, or at all by any given system.

An image differencing technique has been implemented in this change detection study. According to recent research by Coppin & Bauer (1996), image differencing appears to perform generally better than other methods of change detection; and such monitoring techniques based on multispectral satellite data have demonstrated potential as a means to detect, identify, and map changes in forest cover. Image differencing is probably the most widely applied change detection algorithm for a variety of geographical environments (Singh, 1989). It involves subtracting one date of imagery from a second date that has been precisely registered to the first.

This study focuses on previous research by Maus et al (1992), Doak and Lackey (1993), and Green et al (1994) who were successful at detecting, delineating, and classifying forest canopy changes employing an image differencing technique with multitemporal Landsat TM images. Research by Maus et al (1992) focused on the assessment of rural vegetation change on both private and federal land in the Fristoe Unit of the Mark Twain National Forest in southern Missouri. The primary interest in the change detection study by Doak & Lackey (1993) and Green et al (1994) was to identify urban natural areas that had sustained significant loss of vegetation in Portland, Oregon, primarily through timber harvesting and development related activities.

Although each project was focused in different landscape settings (urban verses rural), image differencing change detection methods were utilized and determined to produce accurate results. Specifically, each analyst derived difference files by employing a Band 7 subtraction from Landsat Thematic Mapper imagery. Such derived TM classifications can greatly aid in the monitoring of forest related activities over time and the results can be readily integrated into a GIS. Assessing change in land cover is important both as a monitoring tool and as a means of efficiently updating digital databases. Further discussion of the Band 7 subtraction image differencing technique employed in this study is presented in Chapter Four, Methodology.

A Review of Local Environmental Research (return to top)

Anthropogenic altercations of the natural landscape by means of urbanization, agriculture, and forestry have been a continuous and increasing process for millennia. Regions of natural vegetation and land cover are removed and replaced with numerous human-managed systems of altered structure. The resulting land use and land cover patterns are composed of both the natural and human-developed environments. These altercations and subsequent patterns of the rural landscape in Northwest Arkansas have been the focal point of much research (Scott & Hoffer, 1995, Wall, 1996, Scott et al, 1992, and Fye, 1999). Much of this research indicates that many current land use practices in Northwest Arkansas are occurring on environmentally sensitive lands.

The rural landscape is a mosaic of natural and human-managed patches that vary in size, shape, and arrangement (Turner, 1990). The shifting patterns of land use and land cover can result in stressed relationships between remaining natural areas and introduced human-managed systems. Shifting patterns of land use and land cover pose a threat of disturbance to the biosphere and has the potential to adversely affect the biodiversity of local ecological communities. The preservation of forest, soil, water, and biodiversity resources is the underlining theme as well as an important and essential component of current environmental research.

Scott and Hofer (1995) focused on the spatial and temporal analysis of the morphological and land use characteristics of the Buffalo river watershed employing remote sensing and GIS modeling techniques. Portions of this watershed reside in Madison, Newton, and Boone Counties which border Carroll County to the southwest, south, and east respectively. The results of their analysis indicate that in nine counties in northwest Arkansas over the 27 year period of study from 1965 to 1992, a considerable conversion from forest to pasture occurred on steeper slopes and on some of the better quality soils in the watershed. The slope categories implemented in their research were also incorporated in this study.

Wall (1996) further reiterated the analysis of Scott and Hofer (1995) in Searcy County, Arkansas and linked the impetus for the recent forest to pasture conversion to increased small-scale cattle production by private landowners. Wall suggests the potential economic and environmental unsustainability of such practices by stating, "increasing cattle production follows previous patterns in terms of its lack of long term economic viability and environmentally destructive practices." Wall also provides additional, localized insight on a variety of reasons why small-scale farmers and ranchers continue to deforest their lands and engage in such ‘questionable’, agricultural practices.

Additional research by Scott et al (1992) analyzed the geographical and statistical relationships between landscape parameters and water quality indices in the Muddy Fork Watershed in Washington County, Arkansas also by utilizing remote sensing and GIS modeling techniques. The impetus of their research stems from the fact that this particular watershed contains many of the key elements involved in rural water quality issues: a high density of farm animals located on well-drained soils over carbonate terrain. Such terrain is considered to be highly susceptible to infiltration of pollutants and can often result in non-point source contamination of water resources. Similar conditions and terrain also occur in Carroll County as well as many areas of Northwest Arkansas. The results of their research indicated that increased levels of nitrates and phosphorus contamination in both ground water and soil could be linked to pastures characterized by well-drained soils and proximity to chicken houses as a result of applied chicken litter for soil amendment or fertilizer.

Research by Fye (1999) utilized GIS modeling techniques to analyze the rate and progression of change in forest characteristics over a 30 year period from 1968 to 1995 in 16 northwest Arkansas Counties, including Carroll County. Overall, Fye’s results did not indicate significant changes in the total amount of forested lands over the combined study area. However, Fye’s research did indicate that there has been some loss of forest and changes in forest type, age, density, and coverage over the last three decades. Fye also suggests that such changes may lead to the fragmentation of forests, altered wildlife habitat, changes in microclimate, soil erosion, and degradation of forest integrity.

An additional research project that has been well documented is the Arkansas GAP Analysis Project (Smith et al, 1998). This nationwide project was developed by biologists to map biodiversity and identify ‘gaps’ in its protection. Several biological variables, such as vegetation, vertebrate distributions, and endangered species, are entered into a GIS, from which biodiversity maps are generated, which are then overlain with land management and ownership data. Unprotected components of biodiversity are identified as ‘gaps’. Specifically, Arkansas GAP Analysis maps were created from Landsat Thematic Mapper satellite imagery and GIS maps of geology, topography, soil and other physical features, and databases of species occurrence and habitat characteristics. Such a database for the entire State of Arkansas contains a wealth of important and viable, environmental data and information from which additional research may be compared.

As this research illustrates, many current land use practices in northwest Arkansas have the potential to adversely effect and degrade the environment with respect to forest, soil, water, and biodiversity resources. Issues concerning the preservation of these natural resources will continue to take increasing importance. The occurrence of shifting patterns of land use and land cover in northwest Arkansas is not altogether dissimilar to that of the more publicized regions of the world. Especially changes occurring in a zone of agriculture and forest interface where the general boundary between agricultural pastures and forests is often fuzzy. Although the expanse and scale of forest conversion may not be comparable to that of the Brazilian Amazon or the tropical forests of Southeast Asia, that fact that similar, shifting patterns of land use and land cover are occurring is evident. While many of these issues concerning the degradation of the environment are of concern worldwide, this research hopes to bring some of the focus to our own backyards.

Chapter 4 - Methodology

Center for Advanced Spatial Technologies

University of Arkansas

Literature Review

Other LULC Change in Carroll County areas: