1. Research ObjectiveThe primary objective is to ascertain what the most efficient, effective, and accurate available remote sensing data are for wildlife habitat and land cover mapping in rural communities. Multiple resolution data are available at a broad scale of costs, financial and time allocation, for users (Key et al., 2001). However, identifying cost versus effectiveness, labor versus accuracy, and processing time versus technical capabilities are still not determined for many common rural planning and management procedures. ASTER image Remotely sensed data are captured on multiple remote sensing platforms at different spatial and temporal resolutions. The capability of the sensing device coupled with the altitude and temporal frequency at which data are acquired can impact results and accuracy of any investigation. The time and cost of fieldwork is usually a trade-off for the cost of high spatial and temporal resolution data. Finer resolution data provide an alternative to labor intensive field monitoring (Wulder et al. 2004). These data are usually purchased from private or international sources and require increased computer processing abilities and storage space. However, the use of finer spatial and spectral resolution data are often cost and resource prohibitive, especially by many rural users and agencies. The basis for this work will be to determine the cost, effectiveness, efficiency, and reproduction from four different remote sensing platforms, Landsat, ASTER, IKONOS, and aerial photography. Land cover data can then be used to improve planning, management and overall wildlife habitat mapping on the Yakama Nation Reservation, Washington. |
2. Research Methods NAIP image Increasingly, different data sources are available through geospatial clearinghouse websites. In this module, we will draw upon readily available and free digital data sources and compare them with finer resolution data. Landsat, ASTER, air photos, and digital ortho-photoquadrangles (DOQQs) digital data are available as fairly inexpensive data. These will be compared using different variables to the finer resolution data purchased from IKONOS. Differences in data cost, methods, efficiency, and accuracy for land cover mapping will be assessed. The results from this data set may be used for restoration, planning, or improved wildlife habitat management (Wulder et al., 2004; Tuxen 2007). A matrix with indicators determining the amount of time, ease, and spatial and temporal accuracy will be filled out during the process and then made available as part of the final results of this research. The different types of digital data will be analyzed using ERDAS Imagine software and integrated with ESRI ArcMap. All technicians involved in the project will be required to keep a detailed record for describing processing methods. The comparison of results from the different image sources will demonstrate and document the level of quality vs. cost for each data type. All information will be made available as a web-based educational and image database resource so rural users can better decide which data and resolutions may be the most effective. |
3. Study Area |
Study Area: The Yakama Ecological Restoration Project Lower Satus restoration site The intensive study area that is selected for this module is within part of an ecological restoration project located near Yakima, Washington. The Yakama Ecological Restoration project began in 1992. It is designed to protect, restore, and manage floodplain lands along anadromous fish-bearing waterways in the agricultural portion of the Yakama Reservation. The goals of the project are to protect and restore at least 27,000 acres of floodplain habitats, integrating fish and wildlife restoration with Yakama cultural restoration. Lands are protected through purchase, lease and easement. Large, contiguous areas of riparian and agricultural areas are targeted for restoration. The restoration of floodplain ecological function is the primary means of achieving these goals. Channels and wetlands are reconnected, and the lands once removed from the floodplain are allowed to flood again. The Yakama Ecological Restoration project goes beyond simple stream buffers in that the whole floodplain is restored. Hydrologic reconnection in this manner allows for natural reintroduction of native wetland and riparian plants. Upland areas are reseeded to native grasses. Funding by the BPA covers the land protection, some of the monitoring, and the land management. Large restoration projects are funded through a variety of federal and non-Federal sources. (Hames, 2008) Mapping land cover is one initial step for assessing target zones for vegetation restoration and wildlife habitat planning. The selection of an intensive study area in a riparian zone north of Wapato, WA was determined to be a conducive site for the focus of this module due to ecological parameters. |
5. Intensive Study Area The boundaries of the study area used for each data product. This subset is from an Ikonos 2008 Image 4. Tutorial AssumptionsThis tutorial assumes that the user has internet access, basic GIS experience, and access to ERDAS Imagine remote sensing software. There are other GIS and digital image processing software packages available. MultiSpec is a freeware program that can provide image processing and spectral analysis if a licensed program is not available. However, the steps and graphics in this module are based upon image processing techniques undertaken with ERDAS Imagine 9.3.ERDAS Imagine is a geospatial data program that allows for advanced processing, data analysis, and data presentation. In this module, ESRI ArcMap is also utilized in tandem with ERDAS for vector analysis and data presentation. One of the essential factors that determines the The acquisition of the data used in this module is through Federal agencies or through private companies. The websites for acquisition are provided on the module with hyperlinks to the agency for the specified data source. Many of the data sites require a formal registration with an email address and some general user information. A tutorial on remote sensing may be necessary for those that are unfamiliar with the terms, techniques, and capabilities of multispectral imagery. A useful tutorial online is the NASA sponsored Dr. Nicholas Short's Remote Sensing Tutorial.Working with digital data also requires knowing and understanding map projections, coordinates and grid systems, and datums. One of the most comprehensive tutorials for this is available from Peter Dana,s website at the The Geographer's Craft. 5.Introduction to Remote Sensing TermsSensors detect electromagnetic radiation from different sensing platforms (ie., airplanes, satellites, or shuttles). The sensing device and technologies for acquiring the remote data often are characterized by their resolutions. Therefore, remote sensing analysts frequently refer to different types of resolutions to describe geospatial data. There are four principal resolutions:A satellite or digital image is often known as a scene or a 'footprint' for the area captured by the sensing device. The selection of a scene often depends upon the study question and general knowledge of environmental parameters in the scene area. For example, if the detection and mapping of a particular vegetation type is desired then it is also necessary to know when the 'green-up' and growth cycle is for that species. Other environmental variables, such as season, time of day, amount of cloud cover are also very influential for determining how and what scene to select. If cloud cover obscures a portion of the scene that is under investigation, then finding another date of imagery becomes necessary. Ancillary data, such as digital elevation models (DEMs), derived from topographic data, LIDAR, or SRTM (Shuttle Radar Topography Mission), provide another geospatial layer to assist with analyses. State agencies concerned with natural resources often make state and county level geospatial datasets available on their websites. In the State of Washington, the Department of Natural Resources and the Department of Ecology have GIS data clearinghouses with watershed specific data. 6. Land Use and Land CoverLand cover mapping can proceed through in situ field methodologies or through automated digital procedures. The former is dependent upon a visit to the field site, while the latter benefits and is improved with field data. Fieldwork or ground-truth is an essential component of land cover research. In situ field knowledge supports image analysis by providing the analyst with data about the landscape under investigation. Transferring field knowledge to the synoptic coverage offered in satellite imagery strengthens the analysis and conclusions that can be drawn and made from imagery. A land-cover classification requires determining land cover class categories that can be defined as fairly homogenous. This includes the various vegetative land cover communities found in a study area that can be considered mappable units.A mappable unit may be determined by the resolution of the data. For instance, in a Landsat 7 ETM+ scene, 1 pixel is approximately 30 meters. Therefore, a mappable unit within remote sensing oftentimes corresponds to the resolution of the imagery. For this reason, there may be different land cover classifications that correspond to increasing levels of detail of resolution. The decision process of defining land cover classes can go through iterations in order to achieve optimal categories to satisfy research questions, stakeholder inputs, and/or floristic inventories. Land cover classes also may coincide with other classifications that have been established. The Anderson Land Use Classification (Anderson et al., 1976) is one of the first standardized hierarchical land cover classifications devised for remote sensing analysis. The classification system is nested hierarchical, meaning that at the primary level it starts with a very general name for a land cover type (for example, Forest) and then numerically moves to increasingly more specific land cover types (for example, Cottonwood). |
7. Land Use and Land Cover Categories for Module
|
8. Data Products Used in this Module
|
|
 |
Comparative Assessment of Multiplatform and Multistage Remote Sensing for Rural Habitat Assessment Initiatives Dr. Jennifer Lipton, Ben Sainsbury - Central Washington University
|
» Home » Aerial Photos
