Jonathan D. Jordan

Enhancing Land Cover Mapping using Landsat Derived Surface Temperature and NDVI

Accurate representation of the physical features of the land covering the watershed is required to understand hydrologic processes, since the extent and type of watershed cover affects the movement of water in the hydrologic cycle. Satellite imagery from Landsat and other satellites provide land cover information with high temporal frequency and spatial accuracy. The utility of these data for understanding hydrologic processes depends on how accurately they are interpreted and mapped. This paper deals with the use of the surface temperatures derived from the thermal band of Landsat images in combination with Normalized Difference Vegetation Index (NDVI) to improve the land cover mapping beyond conventional unsupervised classification of the visible/short wave bands. The study was done on Econlockhatchee River basin, part of the Upper St. Johns River basin in central Florida, Landsat images from 1984, 1994 and 2000 were obtained, and unsupervised classification was conducted. Surface temperatures and NDVI were computed and scatter diagrams were drawn to identify the spectral classes resulting from the unsupervised classification. The results show that spectral signatures in the thermal and near-infrared spectrum are very helpful to improve land cover mapping and discriminate land cover classes. In particular, the surface temperatures and NDVI were useful for delineating boundaries between wetlands and water bodies and between upland and wetland forests.

Landsat Imagery in Runoff Volume Estimation

Remote sensing has been used in the field of hydrologic modeling as a source of data to address the spatial variability of hydrologic processes including storm runoff. Although remote sensing does not directly provide a means of runoff estimation, it can provide quantitative land cover information of suitable spatial resolution that is extremely valuable for model inputs. The USDA-NRCS runoff curve number method is employed in this study to assess the spatial and temporal variability of storm runoff in relation to historical land use changes. The study area was the S-65A sub-basin of the Kissimmee River basin in south Florida. Land use data were obtained from Landsat images for the years 1980, 1990 and 2000, which include a period of development and a period of wetland restoration. Unsupervised classification was performed to assess the changes in land cover, with augmentation from Landsat-derived surface temperatures (for 1990 and 2000 images) and derived vegetation indices. Using the runoff curve number techniques, runoff volumes were determined and comparisons were made. Results show that the temporal and spatial variability of runoff volume resulting from changes in land cover can be readily determined from Landsat images.

Identification of Effective Wavebands of Aerial Hyperspectral Imagery for Wetland Vegetation Mapping in Fort. Drum Marsh, Florida

Aerial Hyperspectral imaging techniques are considered to be much more accurate due to the high spectral resolutions, but the sensors and relevant techniques require higher cost for the users. The goal of this paper was to identify the most effective wavebands of hyperspectral images for the wetland vegetation mapping. A aerial hyperspectral imaging system with a imager of 64 wavebands ranging from 399.2 nm to 920.5 nm and the pixel size of 1 meter was utilized in Ft. Drum Marsh, Upper St. Johns River Basin, Florida. The Jeffries-Matusita distance was used to calculate the measurement of separability with consideration of reducing wavebands to 3, 4, and 5, respectively. The reduced-waveband images proved successfully in the mapping of wetland vegetation species with above 85%, identical pixels as the generated from the full-waveband image. Therefore, it is suggested to utilize those few effective narrow-length wavebands to build a budget and effective aerial imaging systems for some specific but frequently imaging mission.

Airborne hyperspectral imaging for sensing phosphorus concentration in the Lake Okeechobee drainage basin

Eutrophication disturbs the ecological balance in the Lake Okeechobee due to high concentration of phosphorus emanated from the regions in the lakes drainage basin. Ability of measuring phosphorus (P) concentrations of water in the Lake Okeechobee itself is very important. Furthermore, monitoring P in its drainage basins is crucial in order to find the cause of P loading and contributing regions. Also, inexpensive real-time sensing capability for a large area in a short time would help scientist, government agents, and civilians to understand the causes, spot the high-risk areas, and develop management practices for restoring the natural equilibrium. In order to measure P concentrations in the Lake Okeechobee drainage basin, airborne hyperspectral images were taken from five representative target sites by deploying a modified queen air twin engine aircraft. Each flight line covered a swath of approximately 365 m wide. Spatial resolution was about 1 m. Spectral range covered was between 412.65 and 991.82 nm with an approximate of 5 nm spectral resolution. Ground truthing was conducted to collect soil and vegetation samples, GPS coordinates of each location, and reflectance measurement of each sample. On the ground, spectral reflectance was measured using a handheld spectrometer in 400-2500 nm. The samples were sent to a laboratory for chemical analysis. Also diffuse reflectance of the samples was measured in a laboratory setting using a spectrophotometer with an integrating sphere. Images were geocorrected and rectified to reduce geometric effect. Calibration of images was conducted to obtain actual reflectance of the target area. Score, SAM (Spectral Angle Mapping), SFF (Spectral Feature Fitting) were computed for spectral matching with image derived spectral library.

Multispectral Image Analysis for Phosphorus Measurement in Bahia Grass

Current phosphorus (P) measurement for vegetation samples requires sample collection, processing, and analysis in a laboratory. This method is time consuming, costly, labor intensive, and restricted to small area. Remote P sensing using multispectral imaging may contribute the rapid, inexpensive, and extensive measurement of P in terrestrial areas. An airborne multispectral image of Bahia grass (Paspalum notatum) in the Plant Science Research and Education Unit, Citra, Florida was acquired in order to study P concentration in vegetation with a spatial resolution of 0.14 m in four spectral bands (Red@610-660 nm, green@535-585 nm, blue@430-490 nm, and NIR@835-885 nm). A field trial was set up in the imaging site in order to obtain P calibration. This Plot had 24 subplots with different P concentration. P application rates were 0, 20, 40, 80, 120, and 160 lb/acre. Soil and Bahia grass samples were collected for six consecutive weeks. Reflectance of wet and dry soil and vegetation was measured in a laboratory using UV-VIS-NIR spectroscopy. Geometric and radiometric corrections were performed on the multispectral imagery, which was analyzed for the relationships with P concentration.

DEM-GIS Based Spatially Distributed Storm Runoff Response Study Using Remotely-sensed Data

The availability of spatial information from remote sensing, and the ability of geographic information system (GIS) to handle such data efficiently, have contributed much to the understanding of watershed runoff response and land cover changes. Distributed watershed models are commonly used to investigate rainfall-runoff processes. These watershed models require topographic drainage information, such as watershed boundaries and drainage divides which can be derived from Digital Elevation Models (DEMs). Spatially distributed runoff was estimated for the Simms creek in the St. John’s River Water Management District (SJRWMD), Florida using the United States Department of Agriculture, Natural Resources Conservation Service-Curve Number (USDA-NRCS-CN) method. Land use data from Digital Orthophoto Quarter Quadrangles (DOQQ) and Landsat Enhanced Thematic Mapper Plus (ETM+) for 1990, 1995 and 2000 were used to estimate spatially distributed curve numbers. A DEM-GIS based runoff routing technique based on 1-D kinematic wave flow was made to generate hydrographs based on travel time to the watershed outlet using spatially distributed data. Comparison was made between the observed and predicted runoff hydrographs and three existing models: the Time-Area method, the Snyder unit hydrograph model and TOPMODEL, a semi-distributed saturation excess model using, 17 storm events from 1990, 1995, 1999 and 2000. The results indicate that the distributed travel time method can accurately predict runoff response for large isolated storms and the travel time approach performed better with higher efficiencies than the other two models. The Time-Area method performed better than the other models with average efficiency of 0.7.

Runoff Change Study Using Remote Sensing and GIS in Southern Taiwan

Taiwan has been experiencing fast economic development and population growth for the past twenty years. This development has put pressure on the agriculture and overall hydrology of the nation. To cite but few, peak runoff increases significantly in the urbanized area due to the increase of runoff coefficient which was caused by the over-development of neighboring farmland, the construction of new houses, roads, factories, nursery and plastic-mulched farms. Furthermore, the drainage system has not been upgraded accordingly and, as a result, poor drainage and severe floods have occurred. A study employing remote sensing for obtaining periodic regional updates of runoff model parameters through land cover analysis, together with geographic information system (GIS) for handling and performing the image processing and runoff curve number estimation, was performed in two, one static (Mei-Nong) and the other developing (Niao-Song) basins in Kaohsiung, southern Taiwan. Landsat 30-m resolution imagery from 1990, 1995, and 2000 was processed to land-cover maps. Ground truth data was collected from field survey and historical aerial photos. GIS analysis of this land-cover, together with soil map data, was used to estimate spatially-distributed United States Department of Agriculture, Natural Resources Conservation Service-Curve Number (USDA-NRCS-CN) on a 30-m grid, and to compute runoff depths. Results of the study indicate that, integrating remote sensing and GIS for tracking the land cover change hence runoff response was helpful for understanding the trends and possible impacts of the dynamics of the change in land use on the overall hydrology of the two study areas. Niao-Song basin has experienced changes in land cover hence runoff response more than the Mei-Nong basin. Over the 10 years of study, the runoff depth has changed much attributed to an increase in residential areas and reduced agricultural areas. The results of the study will be useful in planning, managing and anticipating the potential change in the runoff response of the irrigation areas as a result of possible changes in land use.

Wetland Quality Assessment Using Landsat Imagery and GIS

Reflectance characteristics obtained from Landsat TM imagery were used to assess the condition of cypress (Taxodium sp.) wetlands in West-central Florida. The health classifications, or quality rankings, were based on the qualitative assessment by the Southwest Florida water Management District of 184 cypress wetlands in a karst region north of Tampa Bay. Landsat images from 1992 and 1994 were used in combination with vector GIS Land cover data to identify some 6000 cypress wetlands in 2000 square km region. Differences in the mid-infrared bands were observed among the reflectance signatures of the 184 wetlands of different quality rankings, which allowed the classification of all 6000 wetlands as severely stressed, moderately stressed or good condition. Classifications were preformed using GIS databases from two different sources, and the results were compared.

Assessing Spatial and Temporal Changes of Runoff Response Using Remote Sensing and GIS

The availability of spatial information from remote sensing and the ability of GIS to handle such data efficiently has contributed much to understanding runoff response and microclimate changes of watershed more accurately. Runoff Curve Number (CN) was determined based on the factors of hydrologic soil group, land use, land treatment, and hydrologic conditions. Spatially distributed runoff using the USDA-NRCS-CN (curve number) method was estimated for Etonia creek and Econlockhatchee River sub-basins (Econ) in St. John’s River Water Management District (SJRWMD), Florida. Land use data from Digital Orthophoto Quarter Quadrangles (DOQQ) and Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper (ETM+) were considered for 1973, 1984, 1990, 1995 and 2000. Soils data were obtained from USDA STATSGO database. Land use changes and the resulting runoff changes were analyzed for the 1973 to 2000 time period. The results indicate that this approach can explain the spatial variability of runoff response efficiently. Grid statistics were computed to see the temporal variability of the runoff depth. Maps of coefficient variation indicate areas of land use change and hence change in runoff response. The Econ sub-basin has shown a larger increase in urban development and hence a larger increase in runoff volume than Etonia sub-basin.