1999 Spotlight on Graduate and Undergraduate Research at ESF
Environmental Resource and Forest Engineering Abstracts

Contents

RELIABLE AND FAST 0N-THE-SPOT MEASUREMENT DEVICE TO ASSESS WOOD MEMBERS DEFLECTION UNDER TEST CONDITIONS. Matthew Annabel, Anthony Browne, Gerald Fielding, William Grove, Miguel Reynero, and Gunther Schnorr.

CLASSIFICATION OF ONONDAGA COUNTY USING A LANDSAT THEMATIC MAPPER IMAGE. Jennifer Barber, Amy Becker, Chris Florack, Trevis Gigliotti, Soogeong Myeong, Michael Plastino, Siripun Taweesuk, Cheng Zhu, and Paul Hopkins.

ACCURACY ASSESSMENT OF SPATIAL MODELS. Jennifer Barber and Paul Hopkins.

A FIELD AND MODEL STUDY OF GALLEY SYSTEMS IN WATERSHEDS OF THE NEW YORK CITY DRINKING WATER SYSTEM. Kelly Bazukiewicz and James M. Hassett.

FUZZY CLASSIFICATION OF REMOTELY-SENSED IMAGERY. Amy M. Becker.

BIOFILTRATION AND VERMICOMPOSTING IN OFFRS®, ON-FARM FERTILIZATION RECOVERY SYSTEM. Jerose, Brian, F. Joseph Van Schaick, Richard McClimans.

TRANSPORT OF NUTRIENTS FROM ON-SITE TREATMENT DEVICES TO FENS: PRELIMINARY PLUME CHARACTERIZATION AND EVALUATION. Monica Coyne, Donald Siegel, David Coler, Jeffrey McDonnell and James M. Hassett.>

THE EFFECT OF ALTITUDE, VIEW ANGLE AND SUN ANGLE, VIEW ANGLE CORRECTION PROCEDURES AND THE ATMOSPHERE ON VEGETATIVE INDICES DERIVED FROM DIGITAL IMAGERY. M. J. Duggin, G. J. Kinn, J. K. Muller, S. Myeong, M. Yavuz, C. Florack and J. Walton.

FOREST MAPPING WITH HIGH RESOLUTION DIGITAL IMAGERY. Christopher Florack, Trevis Gigliotti, Lindi Quackenbush, Dr. Paul Hopkins.

GPS PRECISION AND EFFICIENCY UNDER VARIABLE FOREST TYPE AND CROWN DENSITIES. Elif Karakurt and Robert H. Brock.

REMOVING CESIUM-137 FROM CHERNOBYL-CONTAMINATED MILK: AN UNDERGRADUATE RESEARCH EXPERIENCE AT ARGONNE NATIONAL LABORATORY. Eric Moosbrugger and James M. Hassett

A BRIDGE SCOUR ANALYSIS OF THE NYS ROUTE 13 BRIDGE OVER FALL CREEK, TOMPKINS COUNTY, NY. Michael Petrucco, Otto Herr, Andrew Murphy and James M. Hassett.

AGGREGATION OF SPATIAL DATA INDICATORS: A CASE STUDY OF THE ENVIRONMENTAL PROTECTION AGENCY INDEX OF WATERSHED INDICATORS. Michael Plastino.

SATELLITE IMAGERY ANALYSIS TO ASSESS AFTER-FIRE CHANGES ON GRASSLANDS IN CORDOBA, ARGENTINA. Miguel Angel Reynero and Paul Hopkins.

DESIGN OF A HYDRAULIC FLUME FOR THE RENOVATED BAKER LAB. Christopher Somerlot, Jessica Ross, Edward Hattenberger and James M. Hassett.

SEDIMENTATION BASIN EFFICIENCY: AN UNDERGRADUATE NATIONAL SCIENCE FOUNDATION RESEARCH EXPERIENCE AT THE VIRGINIA POLYTECHNIC INSTITUTE. Scott Stoller and James M. Hassett.

ANALYSIS OF SMALL SCALE HYDROELECTRIC GENERATION POTENTIAL AT A SITE IN NORTHEASTERN PENNSYLVANIA. Scott Stoller, John Klamut and James M. Hassett.

COMPARISON OF SPATIAL AND NON-SPATIAL STATISTICAL MODELS TO DESCRIBE LEAD CONCENTRATIONS AT SKEET AND TRAP RANGES. William Thayer and James M. Hassett.

APPLICATION OF SPATIAL MODELS TO DESCRIBE LEAD CONCENTRATIONS THE GENESIS OF THE 1998 LABOR DAY STORM. Jeffrey Waldstreicher and James M. Hassett.

EXPERIMENTS IN CLASSIFYING NORTHEASTERN FORESTS USING TEXTURE ANALYSIS OF JERS-1 RADAR IMAGERY AND GPS REFERENCE DATA. Cheng Zhu, Soojeong Myeong, Kevin Riordon, William Johnson, and Paul Hopkins.

RELIABLE AND FAST 0N-THE-SPOT MEASUREMENT DEVICE TO ASSESS WOOD MEMBERS DEFLECTION UNDER TEST CONDITIONS. Matthew Annabel, Anthony Browne, Gerald Fielding, William Grove, Miguel Reynero, and Gunther Schnorr, Faculty of Environmental Resources and Forest Engineering, 312 Bray Hall, SUNY - College of Environmental Science and Forestry, Syracuse, NY 13210. Major Advisor: Dr. Robert Brock

     Wood members undergo deflection under the loads of normal working conditions. The range of deflection under different loads has to be known before placing the members into a structure that can be un-stabilized by the member’s movement. Thus, the deflection of wood members needs to be measured in testing conditions before using them.
     We developed a measurement system using a digital camera, and a horizontally and vertically rule-marked panel of wood. The digital camera is installed on a tripod with its optical axis in the same horizontal plane as the member being tested. The rule-marked panel is placed under the member in the same vertical plane. Both conditions reassure the correct setting in an X, Y coordinate system, allowing for photogrammetric measurements. The rule-marked panel appears with no changes in all photos, as ground control points. Two photographs are taken, before and during the test, which are immediately processed with image processing software. The raster format of the images allows us to obtain the scale of the pictures by relating the number of pixels to the actual distance on the rule-marked panel.
     Under these conditions we are able to measure with great accuracy the deflection distance of the member. The complete process takes no more than 15 minutes for each test. The simplicity of the setting allows the use of this system for either a series of tests or only an occasional single test. If a series of tests has to be held, the only delay is on replacing the members that have to be tested. This procedure assures a reliable and fast measurement of a wood member’s deflection under test conditions, without interfering with the basic arrangement of the testing site.

CLASSIFICATION OF ONONDAGA COUNTY USING A LANDSAT THEMATIC MAPPER IMAGE. Jennifer Barber, Amy Becker, Chris Florack, Trevis Gigliotti, Soogeong Myeong, Michael Plastino, Siripun Taweesuk, Cheng Zhu, and Paul Hopkins, Faculty of Environmental Resources and Forest Engineering, 312 Bray Hall, SUNY-College of Environmental Science and Forestry, Syracuse, NY 13210.

     This poster will describe a project to classify Onondaga County landuse using a Landsat-5 Thematic Mapper ™ image from 1992. The landuse classes conform to USGS (Anderson) classification levels I and II with a few modifications for specific conditions. Training sites were compiled for each of the categories using aerial photographs, ground observations, and USGS 7.5" quadrangles. The guided clustering technique (as implemented in ERDAS Imagine) provides a baseline classification. Eight alternative classifications will be assessed for accuracy using GPS data collected using a semi-random stratified sampling approach. Having these accuracy assessments for each of the techniques and the baseline allows for comparison of accuracy and practicality.

ACCURACY ASSESSMENT OF SPATIAL MODELS. Jennifer Barber and Paul Hopkins, Faculty of Environmental Resources and Forest Engineering, 312 Bray Hall, SUNY-College of Environmental Science and Forestry, Syracuse, NY 13210.

     Spatial modeling, a relatively new discipline, has exploded with the increases in computer processing speed and storage space. This rapid development has outstripped the ability of researchers and users in the field to find appropriate methods of accuracy assessment of these models. Researchers are now realizing that methods of map and remote sensing classification accuracy assessment may not always be appropriate. Here the major classes of spatial models are discussed alongside the current assessment techniques for each.

A FIELD AND MODEL STUDY OF GALLEY SYSTEMS IN WATERSHEDS OF THE NEW YORK CITY DRINKING WATER SYSTEM. Kelly Bazukiewicz and James M. Hassett, Faculty of Environmental Resources and Forest Engineering, 312 Bray Hall, SUNY - College of Environmental Science and Forestry, Syracuse, NY 13210.

     A study of galley systems as a wastewater disposal option is being conducted by the U.S. Environmental Protection Agency (EPA), the New York City Department of Environmental Protection (NYCDEP), the New York State Department of Health (NYSDOH) and the State University of New York College of Environmental Science and Forestry (SUNY ESF). The purpose of the study is to asses the effectiveness of galley systems in treating domestic wastewater as compared to conventional absorption trench systems in a surface water supply watershed. Galley systems are constructed of cement chambers that are open on the bottom to allow for seepage of wastewater into the subsurface.
     ASTM monitoring wells were installed at four sites in Putnam County, New York. (just north of New York City). Each site has three downgradient and at least one upgradient well. Biweekly samples have been taken since June 1998 and analyzed for BOD (5-day biochemical oxygen demand), COD (chemical oxygen demand), TKN (total Kjeldahl nitrogen), ammonia-nitrogen and several other water quality parameters. A statistical analysis procedure has been defined and preliminary descriptive statistics have computed. The data indicate that samples from downgradient wells have higher concentrations of the water quality parameters than the upgradient wells.
     A model of one of the sites is under construction using the FEMWATER groundwater-modeling program. FEMWATER is a 3-dimensional finite element model that predicts density dependent flow and transport in variably saturated media. Simulations will be run to compare the galley systems and conventional septic and leach field systems.

FUZZY CLASSIFICATION OF REMOTELY-SENSED IMAGERY. Amy M. Becker, SUNY College of Environmental Science and Forestry, Faculty of Enivronmental Resource and Forest Engineering, 411 Bray Hall Syracuse, NY 13210. Advised by Dr. Paul Hopkins

     Mixed pixels commonly occur in remotely-sensed imagery, especially when dealing with coarse resolution, and are a common problem in land-cover mapping applications. While a mixed pixel is composed of more than one land-class most image classification routines assume "pure" or homogeneous pixels and will allocate the pixel to only one class during classification. This process overly generalizes the data allowing the loss of detail and introduces an incorrect estimation of actual land-class area. By unmixing a pixel into its component parts it is possible to generate a more accurate estimation of different land-cover classes. One way of unmixing these pixels is to use a fuzzy classification technique which allows for varying degrees of association with land-cover categories.
     The objective of this project is to assess the class composition of pixels using a straight forward supervised maximum-likelihood classification verses fuzzy classification. With the fuzzy set method every pixel has a membership value associated with each defined class allowing the possibility of ambiguity regarding the appropriate map class at any location.
     To demonstrate these techniques of classification I will be using a Landsat Thematic Mapper image of the Syracuse Area taken in June of 1987. For the actual classification process Idrisi Version 2 and ERDAS Imagine Version 8.3.1 will be used.
     Although all classification involves a loss of information, good classification not only aims to reduce information loss to a minimum, but by identifying natural groups of individuals that have common characteristics, provides a convenient means of information interpretation. It is hoped that the enhanced ability to evaluate maps resulting from the use of fuzzy sets will improve our understanding of uncertainty in maps and facilitate improved error modeling.

BIOFILTRATION AND VERMICOMPOSTING IN OFFRS®, ON-FARM FERTILIZATION RECOVERY SYSTEM. Jerose, Brian, F. Joseph Van Schaick, Richard McClimans, Department of Environmental, Resource and Forest Engineering, 312 Bray Hall, SUNY Environmental Science and Forestry, Syracuse, NY 13210.

     OFFRS®, the On-Farm Fertilization Recovery System, integrates solid and liquid waste management to obviate land application of manure and mitigate air, soil and water contamination. Manure separation produces 2% solids content liquid manure and reduces liquid volume by 25-30%. Aeration, heat and pH adjustment of liquid manure promotes nitrification, reducing ammonia levels and anaerobic conditions. The biofilter is an enclosed bed of grass (Festuca elatior), on a synthetic substrate known as Residual Biotech Soil (RBS®), inoculated with compost worms (Eisenia foetida). Manure is irrigated on the biofilter at average daily rates of 1.27cm resulting in prolific growth of forage. The biofilter requires 500-750m² per 100 cows. Drains and a geomembrane liner capture clean filtrate for reuse.
     E. foetida ingest fine solids accumulated on the soil surface and produce biologically stable castings that maintain infiltration, aeration, and increase rates of microbial decay through selective feeding habits and burrowing. The RBS® structure imparts high specific surface area and porosity (40 % micropores, 20% macropores, and compacted dry density of less than 0.9 g/cm³) resulting in increased area for microbial habitat, nutrient and cation exchange sites, enhanced root penetration, and resistance to compaction. Harvesting vegetation and castings on the biofilter surface removes nutrients in a desirable form. Research has demonstrated the ability of these treatment combinations to grow vegetation and produce a clean filtrate.
     Thermophilically composted manure solids generate heat captured for greenhouse warming. Reduction of ammonia and volatile solids after 21 days promotes further stabilization as solids are sheet-applied to worm windrows. E. foetida rapidly convert the semi-composted solids to value-added vermicompost.

TRANSPORT OF NUTRIENTS FROM ON-SITE TREATMENT DEVICES TO FENS: PRELIMINARY PLUME CHARACTERIZATION AND EVALUATION. Monica Coyne, Donald Siegel and David Coler, Department of Earth Science, Syracuse University; Jeffrey McDonnell and James M. Hassett, SUNY - College of Environmental Science and Forestry, Syracuse, NY 13210.

     Leachate from on-site treatment devices (septic and galley systems) may degrade nearby wetland ecosystems. We report the preliminary results of filed studies designed to determine the extent to which treatment system discharges have reached waters in three fens in New York State. Two of the fens are located within the Malloryville wetland complex near Ithaca. They were formed within kettle depressions between eskers and kames. The third fen is located in Patterson, NY, adjacent to the Covington Greens condominium complex. This wetland is part of the Great Swamp, which is part of an ancient glacial lake.
     Septic wastes were deposited for several years in a gravel pit about 100 m upgradient of the Malloryville fens, while galley waste leaches into the soil about 30 m immediately upgradient of the Covington Green fen. To define the possible extent of contamination, surface waters and ground waters at variable depths in the peat at each site were sampled in August and November of 1998 and analyzed for specific conductivity, major metals and ligands, bromide, and ammonia. Preliminary results show that one of the Malloryville fens has elevated ammonia and sodium concentrations (9.25 mg/L and 24.4 mg/L, respectively). Sodium, chloride, and specific conductance define narrow plumes in the Covington Green fen. These plumes are oriented perpendicular to the direction that potential septic system solutes would enter the wetland from the galley system. Ammonia concentrations appear to define a localized plume, discharging upwards of the galley margin of the fen from the underlying mineral soil and then diverting in the general direction of the other plume. Additional work is in progress to evaluate the orientation of all plumes in three dimensions and to assess the attenuation of solutes in the wetland systems.

THE EFFECT OF ALTITUDE, VIEW ANGLE AND SUN ANGLE, VIEW ANGLE CORRECTION PROCEDURES AND THE ATMOSPHERE ON VEGETATIVE INDICES DERIVED FROM DIGITAL IMAGERY. M. J. Duggin¹, G. J. Kinn² , J. K. Muller¹, S. Myeong¹, M. Yavuz¹', C. Florack¹ and J. Walton¹, ¹ Department of Environmental Resources and Forest Engineering, 312 Bray Hall, SUNY Division of Engineering, Syracuse NY 13210, e-mail: mjduggin@ix.netcom.com, mjduggin@mailbox.syr.edu; ² WSI/EMERGE, 900 Technology Park, Building 8, 2nd Floor, Billerica, MA 01821, e-mail: gjkinn@task.com.

     In 1997 and in 1998, EMERGE obtained multi-altitude digital image data over several sites, including Oneida County Airport, using a calibrated Kodak DCS 460 CIR camera. In this study, which formed a graduate research project, we examined the raw images obtained for two data sets at each altitude, as well as the same images corrected for the bandpass-dependent lens fall-off with field angle. The digital radiance obtained at each altitude over selected vegetation and over standard targets was used to deduce the normalized difference vegetation index (NDVI). The digital radiance and the NDVI for both the raw and the corrected images were plotted as a function of altitude. Image processing procedures were also used to study the variation of recorded radiance in each bandpass, and the deduced vegetation index with view and illumination geometry. It was possible to study the impact of lens fall-off and the effectiveness of correction for this factor, as well as the effects of the atmosphere, and of cloud shadow and the impact of sun-ground-sensor geometry on the calculated NDVI.

FOREST MAPPING WITH HIGH RESOLUTION DIGITAL IMAGERY. Christopher Florack, Trevis Gigliotti, Lindi Quackenbush, Dr. Paul Hopkins, Faculty of Environmental Resources and Forest Resource Engineering, 312 Bray Hall, SUNY-College of Environmental Science and Forestry, Syracuse, NY 13210

     High-resolution aerial digital frame camera images are being used for agricultural applications. This project focuses on this type of imagery in forestry and developing image processing methods tailored to high resolution data. Conventional approaches to digital image processing were developed for use with satellite imagery and designed for classification of features at a relatively coarser resolution. These techniques fail to make use of the level of detail offered by high-resolution imagery.
     Currently we are exploring two approaches. The first is a traditional technique involving a guided clustering approach to classification. The second approach is to use template matching to delineate individual tree crowns. We use band ratios (eg. NIR/R) as a means of eliminating shadow variations across the canopy. The traditional approach will serve as a baseline to judge the success of the other scheme.

GPS PRECISION AND EFFICIENCY UNDER VARIABLE FOREST TYPE AND CROWN DENSITIES. Elif Karakurt and Robert H. Brock, Faculty of Environmental Resources and Forest Engineering, 312 Bray Hall, SUNY-College of Environmental Science and Forestry, Syracuse, NY 13210.

     The precision of GPS positions for coniferous, deciduous and mixed forest types were studied. Three canopy densities were investigated for each forest type. For each of the nine- (9) plots, GPS files were collected in the winter. GPS files will be collected in the summer for further analysis. Quantitative canopy densities were determined through the analysis of imagery collected at each of the above plots. All GPS files were collected with the same satellite constellation to minimize the PDOP variable. The results are given in terms of the three dimensional precision of the plot locations and the time required collecting a fixed number of GPS positions at each plot.

REMOVING CESIUM-137 FROM CHERNOBYL-CONTAMINATED MILK: AN UNDERGRADUATE RESEARCH EXPERIENCE AT ARGONNE NATIONAL LABORATORY. Eric Moosbrugger and James M. Hassett, Faculty of Environmental Resources and Forest Engineering, 312 Bray Hall, SUNY - College of Environmental Science and Forestry, Syracuse, NY 13210.

     The April 1986 reactor accident at Chernobyl, in the Ukraine, released more than 3 million curies of cesium-137, along with substantial amounts of other radionuclides. The Cs-137 remains the most important health risk because of its long half-life (t_1/2 = 31 yr.) and affinity for milk and other agricultural products. In fact, for the population of the Ukraine, approximately 90% of diet-related exposure to Cs-137 is due to ingestion of milk and milk products.
     I spent the spring 1997 semester at the Argonne National Laboratory as an undergraduate Department of Energy research intern. My research looked at ways of removing Cs-137 from milk products. Specifically, I examined the utility of a magnetic separation technology developed by Selentec, Inc. In this process, magnetic particles, coated with a resinous zeolite (clinoptilolite), are introduced to milk. The resin selectively adsorbs the Cs-137, and magnets remove the particles. Specifically, I investigated variations in the particle sizes and methods used for attaching the zeolite to the magnetic particles. These differences resulted in different amounts of Cs-137 removal from the contaminated milk. Some particle size/attachment method combinations reduced gamma emissions by more than 90%.

A BRIDGE SCOUR ANALYSIS OF THE NYS ROUTE 13 BRIDGE OVER FALL CREEK, TOMPKINS COUNTY, NY. Michael Petrucco, Otto Herr, Andrew Murphy and James M. Hassett, Faculty of Environmental Resources and Forest Engineering, 312 Bray Hall, SUNY - College of Environmental Science and Forestry, Syracuse, NY 13210.

     Floodwaters can remove river sediment from around bridge piers. The resulting scour hole can cause the piers to collapse and result in a catastrophic bridge failure. Bridge pier scour was in fact the cause of the collapse of the NYS Thruway bridge over Schoharie Creek in April 1987. The collapse caused 10 deaths and led to a renewed interest in developing models to predict bridge scour at existing bridges.
     We applied two scour prediction methods to the case of the New York State Route 13 bridge over Fall Creek, located in Tompkins County, NY. We obtained river sediments from the vicinity of the bridge pier and analyzed them for particle and sediment characteristics. We used an estimate of the one in hundred year flood and output from the Corps of Engineers River Analysis System software to determine approach flow depth and velocity. We used bridge plans obtained from the NYS Department of Transportation to model bridge pier and channel geometry, which we verified by field observations.
     Our results show that the bridge is at risk for failure due to bridge pier scour. Our calculations predict a scour depth of 11 to 15 feet, depending on the method used. Our results are consistent with NYSDOT calculations; the bridge is under a flood watch and is scheduled for replacement in the near future.

AGGREGATION OF SPATIAL DATA INDICATORS: A CASE STUDY OF THE ENVIRONMENTAL PROTECTION AGENCY INDEX OF WATERSHED INDICATORS. Michael Plastino, Faculty of Environmental Resources and Forest Engineering, Mapping Sciences Program, 312 Bray Hall, SUNY-College of Environmental Science and Forestry, Syracuse, NY 13210. Faculty Advisor: Dr. Paul F. Hopkins.

     Aggregation of spatial data indicators has become increasingly common with the growing spatial data processing power of geographic information systems (GIS). Using the Environmental Protection Agency (EPA) Index of Watershed Indicators (IWI) as a case study, aggregation methods and weighting schemes are explored and evaluated for accuracy in determining the overall health of water in the 2,111 watersheds of the continental U.S. IWI indicator data is aggregated with the IWI summation, simple summation, scaled summation, and scaled multiplication methods and weighted to account for data distribution, correlation, quality, and a combination of the three. Aggregation methods / weighting schemes are tested using a set of 100 sample watersheds. Results are compared with two evaluation data sets. The expert-weighted set is an evaluation of the 100 sample watersheds using expert weights derived from the average indicator priority rankings from 9 experts. The average expert watershed assessment set is the average of three expert watershed assessments, based on indicator parameter values for the 100 sample watersheds.
     The aggregation methods / weighting schemes are evaluated using two methods. One method uses pairwise t-tests to measure the similarity between aggregation methods / weighting schemes results and the two evaluation data sets. The other method compares the averages and standard deviations of aggregation methods / weighting schemes results with the average expert watershed assessment results.
     Evaluation results indicate that various weighting schemes of the scaled summation method fare relatively well in both evaluation methods. The ‘distribution’ weighting scheme also performs well in several evaluations. None of these evaluation results are validated with both evaluation data sets. The success of the scaled summation method and the distribution weighting scheme are therefore not definite, but should be further explored using a more discriminating weighting scheme and more reliable evaluation data sets.

SATELLITE IMAGERY ANALYSIS TO ASSESS AFTER-FIRE CHANGES ON GRASSLANDS IN CORDOBA, ARGENTINA. Miguel Angel Reynero and Paul Hopkins, Faculty of Environmental and Resources Engineering, 410A Bray Hall, SUNY-College of Environmental Science and Forestry, Syracuse, NY 13210.

     Every year thousands of hectares of grasslands, pastures, forests and tree plantations burn in Cordoba, as well as in the rest of Argentina. Space-borne sensors are a reliable data source from which information about fires on forests, grasslands, and other open or natural areas is produced. That information covers estimation of fire danger potential, early detection, the monitoring of fires occurrences, and after-fire change detection.
     I shall use Landsat TM5 satellite imagery to assess after-fire changes on grasslands located in the South-West of Cordoba Province, Argentina, applying post-classification comparison, Kauth-Thomas transformation and vegetation indices as image analysis methods.
     The study site is part of 300,000 hectares of hills with a mean altitude of 1200 meters above sea level. This area heavily influences the landscape and environment of the plains extended toward South and East, due to sediments, salts and water originating in the hills. The main economic activities are cattle raising, and forestry, affected both by fire because of loss of forage, fences, trees, and death of domestic and wild animals. Fire induced changes in vegetation growth, biomass production, pattern of vegetation patches, and species composition of plant communities, are directly related to these environmental and economical effects. Changes detected in the burnt areas of the study site will support the planning of restoration and management of similar situations in the region.

DESIGN OF A HYDRAULIC FLUME FOR THE RENOVATED BAKER LAB. Christopher Somerlot, Jessica Ross, Edward Hattenberger and James M. Hassett, Faculty of Environmental Resources and Forest Engineering, 312 Bray Hall, SUNY - College of Environmental Science and Forestry, Syracuse, NY 13210.

     The Faculty of Environmental Resources and Forest Engineering is scheduled to move into the renovated Baker Lab. In anticipation of that move, we have begun to design laboratory facilities for the refurbished building. We are planning a hydraulics lab with a multipurpose flume to illustrate key concepts in fluid mechanics and open channel flow.
     We have a space allocation of about 1000 ft² for the lab. We have designed a variable slope flume, the slope of which is controlled by a hydraulic system controlled by a computer interface. We can therefore illustrate open channel flow characteristics of mild, critical, and steep slopes.
     We have also designed a hydrograph generation system to illustrate relationships between inflow, storage and outflow from a hydrologic system. We have designed the system to be flexible, inexpensive, and a useful tool to help students learn the difficult concepts of open channel flow.

SEDIMENTATION BASIN EFFICIENCY: AN UNDERGRADUATE NATIONAL SCIENCE FOUNDATION RESEARCH EXPERIENCE AT THE VIRGINIA POLYTECHNIC INSTITUTE. Scott Stoller and James M. Hassett, Faculty of Environmental Resources and Forest Engineering, 312 Bray Hall, SUNY - College of Environmental Science and Forestry, Syracuse, NY 13210.

     During the summer of 1998, I participated in an undergraduate research program at the Virginia Polytechnic Institute in Blacksburg, VA. I worked on a project to evaluate the efficiency of sedimentation basins in removing sediments and nutrients from non-point source (NPS) pollution. The specific objectives were to compare the effectiveness of wet vs. dry basins, skimmer vs. riser outlets, and short vs. long dewatering times in terms of both sediment and nutrient removals.
     A series of ponds were constructed which incorporated the features listed above. The ponds were filled with water with constituents similar to agricultural NPS runoff. Influent and effluent were sampled for total suspended solids, nitrate, total Kjeldahl nitrogen, ammonium, total phosphorus, and orthophosphate.
     The ponds showed quite different results for solids removals. The short dewatering time (6 hr), dry ponds with risers removed about 90% of the influent solids, while the long dewatering time (12 hr), wet ponds equipped with skimmers removed about 98% of the influent solids. Nitrate removal was insensitive to pond design factors.
     Our studies show that design features of sedimentation basins influence sediment and nutrient removals, and may lead to a rational design methodology for these important devices.

ANALYSIS OF SMALL SCALE HYDROELECTRIC GENERATION POTENTIAL AT A SITE IN NORTHEASTERN PENNSYLVANIA. Scott Stoller, John Klamut and James M. Hassett, Faculty of Environmental Resources and Forest Engineering, 312 Bray Hall, SUNY - College of Environmental Science and Forestry, Syracuse, NY 13210.

     We analyzed the potential for micro-hydropower generation at a site in northeastern Pennsylvania. We undertook the project at the request of the landowners, and in fulfillment of an engineering design requirement for a senior course in our curriculum.
     We surveyed the site and created a topographic map to further our work. We estimated a flow of 0.11 cubic feet per second (cfs), based on watershed area and actual flow measurements. We specified and provided costs for the major elements of design: pipe, turbine, electrical inverter, batteries, and power line.
     Our design calls for 1280 ft of 4 inch diameter PVC pipe, a one nozzle Pelton Wheel undershot turbine with a maximum output of 600 W, an Exeltech XP-600 DC to AC inverter, and two Trojan L16 batteries.
     Our economic analysis proceeded under the assumption that electrical power from the grid costs $0.1059 per Kilowatt Hour (KWH) and will remain constant. The site can produce 0.360 KW per month, for a monthly savings of $27.50. The analysis showed that the payback period will be very long, and therefore the micro-hydropower alternative is not economically feasible for this site, at least under our design and analysis assumptions.

COMPARISON OF SPATIAL AND NON-SPATIAL STATISTICAL MODELS TO DESCRIBE LEAD CONCENTRATIONS AT SKEET AND TRAP RANGES. William Thayer and James M. Hassett, Faculty of Environmental Resources and Forest Engineering, 312 Bray Hall, SUNY - College of Environmental Science and Forestry, Syracuse, NY 13210.

     Skeet and trap ranges often have high lead concentrations associated with ‘hot spots’, i.e., small areas of high lead concentrations. Skeet and trap ranges thus provide robust test cases to evaluate different methods of estimating a site exposure point concentration (EPC), a value needed to determine if a risk to humans or the environment exists. EPA guidance requires the use of the 95th-percentile upper confidence limit (UCL) of the arithmetic or log-transformed mean of the site contaminant data as the EPC. However, the EPA method sometimes estimates an EPC that is greater than the highest concentration observed in the site data.
     The effect of using spatial and non-spatial models to estimate exposure concentrations at skeet and trap ranges is being investigated using lead concentration data from 5 ranges. Estimates of the 95th UCL calculated in accordance with the EPA method were compared to estimates generated from a bootstrap method. In addition, the bootstrap method was modified to incorporate spatially derived weights for each observation in the data set by means of Vornoi diagrams generated for each site.
     Preliminary results suggest that the use of spatially weighted models produce estimates of the EPC that are very different from and usually lower than the non-spatial methods. Our work suggests that clustering samples in suggested or known areas of high concentrations (the ‘hot spots’) can lead to overly conservative values of the EPC for a site. The overly conservative values may lead to remediation efforts that are needlessly expensive.

APPLICATION OF SPATIAL MODELS TO DESCRIBE LEAD CONCENTRATIONS THE GENESIS OF THE 1998 LABOR DAY STORM. Jeffrey Waldstreicher, National Weather Service, Binghamton, NY, and James M. Hassett, Faculty of Environmental Resources and Forest Engineering, 312 Bray Hall, SUNY - College of Environmental Science and Forestry, Syracuse, NY 13210.

     The 1998 Labor Day storm was unusual in several respects. Early morning severe storms are rare in central New York (exception: the July 15, 1995 event). The maximum wind velocities (in the Syracuse area) of 100-100 mph are also unusual. The extensive damage and loss of three lives also attest to the unusual nature of the storm.
     The storm resulted from an unusual combination of meteorological conditions. First, a layer of warm, moist unstable air was trapped by an atmospheric inversion at mid-altitude. As the warm air moved over central New York, it mixed through the boundary layer, resulting in a rapid increase in surface and dew point temperatures. Second, a cold front was moving into New York from the upper Great Lakes. Finally, a jet stream was moving across southern Ontario into southern Quebec. The combination of factors led to the development of straight line squalls. As the squall line developed, a ‘bow line’ effect caused the winds in the center of the system to become amplified.
     The poster will review the synoptic conditions that led to the storm development, describe the genesis of the ‘bow line’ in terms of a theoretical model, show radar images of the storm, and compare storm conditions to after the fact mesoscale modeling efforts.

EXPERIMENTS IN CLASSIFYING NORTHEASTERN FORESTS USING TEXTURE ANALYSIS OF JERS-1 RADAR IMAGERY AND GPS REFERENCE DATA. Cheng Zhu, Soojeong Myeong, Kevin Riordon, William Johnson, Paul Hopkins, Faculty of Environmental Resource and Forestry Engineering, 312 Bray Hall, SUNY-College of Environmental Science and Forestry, Syracuse, NY 13210

     Radar is an active sensor, operating in the microwave portion of the Electromagnetic spectrum. Radar can penetrate clouds and operate day or night, providing imagery of the surface whatever the weather conditions or the time of day.
     The waves emitted by active sensors travel in phase and interact minimally on their way to the target area. After interaction with the target area, these waves are no longer in phase due to the different distances they travel from targets. Once out of phase, radar waves can interact to produce light and dark pixels known as speckle. Speckle, generally considered to be noise, carries valuable information about the radar imaging system. Texture analysis is used for measuring the spatial variation of the speckle and providing information for image classification.
     To assist classifying different forest types, GPS (Global Position System) reference data were collected for CFI (Continuous Forest Inventory) plots in ESF research properties (Huntington and Heiberg Forests). These GPS referenced data represented the three major forest types of Hardwood, Softwood and Mixed. These GPS data were converted to Arc/Info coverage, then used for creating signature files in ERDAS Imagine 8.3.1.
     Different Texture analysis methods with different window sizes were applied to JERS-1 radar imagery. We texture measures studied were Variance, Skewness, Kurtosis and Mean Euclidean Distance using windows of 3X3, 9X9 and 27X27. Discriminant analysis (SAS) was used to find the optimal combination of texture features.
     Results show that texture analysis of JERS-1 multi-look standard images can distinguish forest and non-forest land cover, but could not separate different forest types. It was also found that radar imagery combined with optical imagery (i.e. Landsat Thematic Mapper) could help classifying different forest types. Further research will use different radar images with the phase information and additional methods for combining radar with optical images.