Remote Sensing, Hydrology, and Field Experiments III Posters
Presiding: M H Cosh, USDA/ARS Hydrology and Remote Sensing Laboratory; J Jacobs, University of New Hampshire; B Mohanty, Texas A&M University
H41A-01 0830h
Satellite Remote Sensing of Pan-arctic Vegetation Productivity, Soil Respiration and net CO2 Exchange Using MODIS and AMSR-E Data
We have developed an approach for regional assessment and monitoring of land-atmosphere carbon dioxide (CO2) exchange, soil heterotrophic respiration (Rh) and vegetation productivity for arctic tundra using global satellite remote sensing at optical and microwave wavelengths. We use C- and X-band brightness temperatures from AMSR-E to extract surface wetness and temperature, and MODIS data to derive land cover, Leaf Area Index (LAI) and Net Primary Production (NPP) information. Calibration and validation activities involve comparisons between satellite remote sensing and tundra CO2 eddy flux tower and biophysical measurement networks and hydro-ecological process model simulations. We analyze spatial and temporal anomalies and environmental drivers of land-atmosphere net CO2 exchange at weekly and annual time steps. Surface soil moisture status and temperature as detected from satellite remote sensing observations are found to be major drivers spatial and temporal patterns of tundra net CO2 exchange and photosynthetic and respiration processes. We also find that satellite microwave measurements are capable of capturing seasonal variations and regional patterns in tundra soil heterotrophic respiration and CO2 exchange, while our ability to extract spatial patterns at the scale of surface heterogeneity is limited by the coarse spatial scale of the satellite remote sensing footprint. Our results also indicate that carbon cycle response to climate change is non-linear and strongly coupled to arctic surface hydrology. This work was performed at The University of Montana and Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.
http://www.ntsg.umt.edu
H41A-02 0830h
Inferring floodplain subsurface flowpaths using remote sensing vegetation indices
Floodplain vegetation directly responds to water availability in the alluvial aquifer. Variation in plant water stress, both spatially and across the growing season, may indicate preferential flowpaths in alluvial floodplains. We use aerial and satellite remote sensing data to monitor change of a semi-arid floodplain through the summer season. Combining several remote sensing instruments in a 2004 campaign, we collected images at multiple spatial scales from 4/2005 to 9/2005. We evaluated several vegetation indices, including Normalized Difference Vegetation Index (NDVI), Soil Adjusted Vegetation Index (SAVI) and Red-edge Stress Vegetation Index (RSVI) and compared them to local floodplain topography. Initial results suggest that RSVI is the most robust index and the least sensitive to atmospheric conditions. Apparent surface reflectance was calculated from each data set and aerial hyperspectral images were convolved to satellite images. By standardizing units in this way, we can directly compare spectra across multiple basins and identify patterns of water stress in riparian vegetation at multiple spatial scales. Corroborating these results with 3-dimensional model outputs and field tracer tests should identify areas where vegetation responds directly to patterns of hydrologic connectivity.
H41A-03 0830h
Modeling Water Clarity Using Aster Multispectral Remote Sensing Data, Lake Tahoe
The decline in water clarity has been an important issue affecting Lake Tahoe since water clarity measurements began in the late 1950's. Traditional estimates of water clarity rely on secchi depth; however turbidity as a proxy for water clarity is gaining popularity with the nephelometer. Both methods rely on point data and unless sufficient measurements are taken the spatial dimensionality of lake water clarity may be grossly underestimated. Previous investigations throughout North America and Europe have successfully correlated satellite data with water clarity. This project explores the utility of remote sensing as an alternative to traditional water clarity measurements at Lake Tahoe. A series of Aster images (Ast 07, reflectance) spanning four years were calibrated to one image (September 2000) using static ground targets. A mask was then applied to each calibrated data set to isolate the lake from the surrounding basin. Aster band 1 and band 3 (520nm - 600nm and 780nm - 860nm respectively) calibrated data sets were classified independently based on absolute reflectance values. Processed image data was then used to create a series of maps depicting lake water reflectance over the entire four year period. Preliminary comparisons of reflectance data with empirical secchi depth and turbidity data indicate a positive correlation. Additional analysis will be performed with the ultimate goal of deriving a numerical correlation between data sets. Reflectance values are seasonally dependent with the highest values consistently occurring in the late spring and early summer months. Variance within the lake is also elevated during this time of year. Seasonal trends and relationships to in situ measurements will be presented at the meeting.
H41A-04 0830h
Improved Forecasting of Spring Snowmelt Runoff in the Ob River Basin Using Satellite-Derived Snow Volumes
In high latitude river basins like the Ob in central Siberia, snowmelt runoff can contribute 75 percent of the annual streamflow. As is the case with many remote basins, in-situ snow data for the Ob basin are sparse, and therefore satellite data must be relied upon for accurate modeling. This current study seeks to improve forecasts of the spring snowmelt runoff through the incorporation of winter satellite derived snow water equivalent into a precipitation runoff model. The snow product used is Scanning Multichannel Microwave Radiometer (SMMR) snow depth imagery with a 0.25 degree resolution for the period 1980-1987. Spatial and temporal biases of the SMMR images are corrected through the development of a set of correction factors using 10-day NSIDC snow depth measurements. The satellite data are used in conjunction with local meteorological forcing and an enhanced version of the original USGS Precipitation Runoff Model to estimate runoff. Results indicate the use of the satellite imagery for both calibration and forcing purposes improves the annual spring snowmelt forecast.
H41A-05 0830h
Moisture Mapping in the Amazon Basin Using Near-nadir and Far-range Backscatter Slopes
The electromagnetic scattering from vegetated targets has both coherent and non-coherent scattering components that correspond to contribution from underlying soil and vegetation canopy, respectively. The non-coherent contribution is higher at far-range incidence angles (θ) than near-nadir θ. The relative contribution of coherent component increases for lower canopy thickness. The relative contribution of coherent and non-coherent scattering components is also effected by the moisture content of canopy and soil. We use normalized radar backscatter (σ°) measurements from ERS Scatterometer (ESCAT), NASA Scatterometer (NSCAT) and Tropical Rainfall Measuring Mission Precipitation Radar (TRMM-PR) to analyze the moisture characteristics and its spatial and temporal variability in the Amazon basin. The amount of coherent scattering is indicated by high near-nadir slopes of σ° θ-response which depends on surface scattering contribution. The rate of decrease in σ° from nadir direction to near nadir directions (θ < 5°) is found to be very sensitive to the vegetation density. The slope drops significantly as the θ increases to 10°, beyond which the non-coherent scattering dominates. For very dense vegetation the near nadir σ° has only non-coherent scattering and hence slope is low. Densely vegetated targets do not exhibit near nadir specular reflection and thus slope increases gradually with increase in θ. We define the Normalized Near-nadir and Far-range Slope Difference (NNFSD) as a discriminant to quantify the relative contribution of coherent and non-coherent scattering. The presence of water bodies (rivers and ponds) significantly increase the specular reflection of electromagnetic waves at near-nadir θ. NNFSD is used to identify large scale water-bodies in the Amazon basin and monitor their spatial and temporal behavior.
H41A-06 0830h
ASSESSMENT USING PASSIVE MICROWAVE REMOTE SENSING AND VADOSE ZONE MODELING
Soil moisture is a fundamental state variable and it varies spatially due to topography, soil, precipitation variability and landuse-landcover, and temporally, due to difference in hydrologic characteristics and controls. Many studies have successfully demonstrated the use of passive microwave remote sensing to assess soil moisture mapping information at a depth of 5 cm. Assessing root zone soil moisture through these remotely sensed data could be achieved by integration of remote sensing and computational modeling. This integrated method may become resourceful solution to the problem for profile soil moisture estimation and its transient behavior. This paper describe a model based on GIS distributed hydrologic modeling coupled with HYDRUS-ET to estimate surface and root zone soil water content using surface soil moisture data from ESTAR (SGP 97) with other topographical and hydro-meteorological data. The model estimated root zone soil moisture for the Little Washita watershed at various depths for the entire SGP97 period. The simulated soil moisture at various depths was compared with time-domain reflectometry (TDR) profile measurements. Overall reasonable agreement was found between model estimations and TDR measurements were the pixel soil profile matched with the point soil profile. The model showed great promise and flexibility to incorporate assimilation techniques, scaling and other hydrological complexities.
H41A-07 0830h
Sitting of Water Harvesting Structures Based on Remote Sensing, GIS and GPS
Harvesting of runoff is critically important for supporting rain fed agriculture in semi-arid areas. In the present study, Remote Sensing,GIS, GPS technologies are used to analyze the data to provide information about suitable sites for water harvesting structures in Khamam District of Andhra Pradesh, in India . The layers used for site suitability index for water harvesting structures are the canopy density, slope map, soil map and tank buffers which give the required indices. The suitability Index map is generated by integrating the above layers; slope map is derived from DEM; forest density map generated by classifying the satellite data obtained from IRS 1C/1D LISS-III, soil map is from ICAR on 1:500000 scale, tank buffers created at 500m distance. Different weightages are assigned as per the importance of the particular variable and spatially modeled using overlay analysis to generate the site suitability index map and rescaled to suit the requirement. The final suitable index values are reclassified into various classes like highly suitable to CD/PT's, moderately suitable and least suitable,etc. Stream network is digitized for the entire study area. Integrating the above said layers for generating site suitability map for water harvesting structures like CD/PT's are proposed along with latitude, and longitude, information. A model is developed for the integrated management of water resources considering the spatial thematic layers and attribute data base. The micro level management is carried out to suggest the locations of the water harvesting structures. The types of structures are designed in accordance with decision rules developed in this model. The model arriving out of this study may be used elsewhere under similar environment.
H41A-08 0830h
Detection of Septic System Performance via Remote Sensing Technologies
Failing and improperly managed septic systems can affect water quality in their environs and cause health problems for individuals or community residents. When unchecked, failing systems can allow disease-causing pathogens to enter groundwater aquifers and pollute surface waters, contaminating drinking water, recreational waterways, and fishing grounds. Early detection of septic system leakage and failure can limit the extent of these problems. External symptoms which occur over an improperly functioning septic system can include lush or greener growth of vegetation, distress of vegetation, excessive soil moisture levels, or pooling of surface effluent. The use of remote sensing technologies coupled with attainable permit records to successfully identify these features could enable the appropriate agencies to target problem areas without extensive field inspection. High-resolution, airborne imagery was identified as having the potential to detect relative changes in soil moisture, to delineate individual leach fields, and to locate effluent discharges into water bodies. In addition, vegetation patterns responding to nutrient-rich effluent and increased soil moisture could be examined using a vegetation index. Both thermal- and color-infrared imagery were acquired for a study area in Jackson County, Mississippi, adjacent to the Gulf of Mexico. Within this coastal neighborhood known to have significant septic system failures, over 50 volunteer residents supplied information regarding the function of their systems and access to their property. Following data collection, regression methods were used to nominate the major indicators of malfunctioning systems. A ranking system for the "level of function" was derived from these analyses. A model was created which inputs data from attainable records and imagery analysis and outputs a predicted level of septic system function. The end product of this research will permit evaluation of septic system performance to be estimated using only easily obtainable data, allowing for minimal effort in the prioritization of problem areas by regulatory agencies.
H41A-09 0830h
Characterization of Footprint-Scale Surface Soil Moisture Distribution Observed during Soil Moisture Experiments in 2004 (SMEX04) Using Gaussian Mixture Model
During the Soil Moisture Experiments in 2004 (SMEX04) in Arizona, isolated precipitation events over the experiment site created patched spatial patterns of surface wetness, which in turn resulted in non-Gaussian distributions of surface soil moisture contents within some satellite footprint-scale fields. Mixtures of two Gaussian distributions are applied to fit the observed distributions of ground-based soil moisture data. Each footprint-scale field is stratified into two sub-regions in order to determine the distribution mixing weights using both the antecedent precipitation index and soil texture. The stratified patterns are compared with stratified high-resolution soil moisture maps from aircraft polarimetric scanning radiometer (PSR) observations. Bayesian information criteria (BIC) from fitting a single Gaussian distribution and fitting mixtures of two Gaussian densities are compared to test the optimality of the models, and uncertainties resulting from both models are compared.