B24B-01
Gray Wave of the Great Transformation: A Satellite View of Urbanization, Climate, and Food Security
Land cover change driven by human activity is profoundly affecting Earth's natural systems with impacts ranging from a loss of biological productivity to changes in atmospheric chemistry and regional and global climate. This change has been so pervasive and progressed so rapidly, compared to natural processes, scientists refer to it as "the great transformation". Urbanization or the 'gray wave' of this transformation is being increasingly recognized as an important process in global climate change. A hallmark of our success as a species, large urban conglomerates do in fact alter their environments so profoundly that the local climate, atmospheric composition, and the basic ecology of the landscape are affected in ways that have consequences to human health and economic well-being. Fortunately we have incredible new tools to observe and understand these processes in ways that can be used to plan and develop enjoyable and sustainable urban places. A suite of Earth observing satellites is making it possible to study the interactions between urbanization, biological processes, and the atmosphere including weather and climate. Using these Earth Observatories we are learning how urban heat islands form and potentially ameliorate them, how urbanization can affect rainfall, pollution, surface water recharge at the local level, and climate and food security globally.
B24B-02
Estimating and Mapping Urban Impervious Surfaces: Reflection on Spectral, Spatial, and Temporal Resolutions
Impervious surface is a key indicator of urban environmental quality and urbanization degree. Therefore, estimation and mapping of impervious surfaces in urban areas has attracted more and more attention recently by using remote sensing digital images. In this paper, satellite images with various spectral, spatial, and temporal resolutions are employed to examine the effects of these remote sensing data characteristics on mapping accuracy of urban impervious surfaces. The study area was the city proper of Indianapolis (Marion County), Indiana, United States. Linear spectral mixture analysis was applied to generate high albedo, low albedo, vegetation, and soil fraction images (endmembers) from the satellite images, and impervious surfaces were then estimated by adding high albedo and low albedo fraction images. A comparison of EO-1 ALI (multispectral) and Hyperion (hyperspectral) images indicates that the Hyperion image was more effective in discerning low albedo surface materials, especially the spectral bands in the mid-infrared region. Linear spectral mixing modeling was found more useful for medium spatial resolution images, such as Landsat TM/ETM+ and ASTER images, due to the existence of a large amount of mixed pixels in the urban areas. The model, however, may not be suitable for high spatial resolution images, such as IKONOS images, because of less influence from the mixing pixel. The shadow problem in the high spatial resolution images, caused by tall buildings and large tree crowns, is a challenge in impervious surface extraction. Alternative image processing algorithms such as decision tree classifier may be more appropriate to achieve high mapping accuracy. For mid-latitude cities, seasonal vegetation phenology has a significant effect on the spectral response of terrestrial features, and therefore, image analysis must take into account of this environmental characteristic. Three ASTER images, acquired on April 5, 2004, June 16, 2001, and October 3, 2000, respectively, were used to test the seasonal sensitivity of impervious surface estimation. Our results indicated that the summer (June) image was better than the spring (April) and the fall (October) ones. The summer image was most appropriate because there was full growth of vegetation, and mapping of impervious surfaces was more effective with contrasting spectral response from green vegetation. The mixing space, based on the four endmembers was perfectly three-dimensional. In contrast, there was significant amount of bare soils/grounds and non-photosynthesis vegetation in the spring and fall images. Plant phenology caused changes in the variance partitioning and impacted the mixing space characterization, leading to the difficulty in the estimation of impervious surfaces. http://www.gis.indstate.edu/
B24B-03
Quantitative Remote Estimation of Land Surface Biophysical Parameters
Characterization of cropland and grassland properties over large areas requires not only the use of imaging remote sensors but also algorithms suitable for robust retrievals of important biophysical properties of vegetation. In this work we offer four techniques for retrieving biophysical characteristics of cropland and grassland using remote sensing. We suggest using the Visible Atmospherically Resistant Vegetation Index (VARI), based on bands only in the visible portion of the spectrum (red and either green or the red edge), for retrieving the fractional cover of green vegetation. The Wide Dynamic Range Vegetation Index (WDRVI), a non- linear transformation of NDVI, allows retrieval of LAI in a wide range of cropland- and grassland-biomass situations. The Green NDVI and Red Edge NDVI were designed to assess the photosynthetically active component of total absorbed photosynthetically active radiation (fAPAR). The Green and Red Edge Chlorophyll Indices are suggested for measuring the total chlorophyll content in a vegetation canopy, which equates for cropland and grassland to gross primary production (GPP). Our results document accurate estimation of fAPAR, fractional cover of green vegetation, LAI, chlorophyll content and GPP using close-range sensing, an airborne imaging spectrometer (AISA-Eagle) and satellite observations. The techniques were calibrated and validated for maize and soybean in agricultural fields under both irrigated and rainfed conditions.
B24B-04
Retrieving Land Surface Properties from MODIS and MISR Albedo Products
We present results from the application of an inversion method conducted using both MODIS and MISR derived broadband visible and near-infrared surface albedo products available during a full seasonal cycle. It addresses complex geophysical scenarios involving snow occurrence in mid and high-latitude evergreen and deciduous forest canopy systems. The occurrence of snow during the winter and spring seasons is based on the analysis of the MODIS snow products which assimilation by our package translates into an adaptation of the prior values, both the maximum likelihood and width of the 2-D probability density functions (PDF), characterizing the background conditions of the forest floor. Our results illustrate the capability of the inversion package to retrieve meaningful land vegetation fluxes and associated model parameters (such as the effective LAI) along the year despite the rather high variability in the input products. As a matter of fact, most of this temporal variability, as well as the small differences between the MODIS and MISR broadband albedos, appear to be mostly captured by the albedo of the forest floor. We will discuss results from applications conducted using MODIS and MISR operational surface albedo products over selected EOS validation sites where some ground-based estimates are available. http://fapar.jrc.ec.europa.eu/
B24B-05
Global sensitivity analysis of Leaf-Canopy radiative transfer Model for analysis and quantification of uncertainties in remote sensed data product generation
Analyzing, quantifying and reporting the uncertainty in remote sensed data products is critical for our understanding of Earth's coupled system. It is the only way in which the uncertainty of further analyses using these data products as inputs can be quantified. Analyzing the source of the data product uncertainties can identify where the models must be improved, or where better input information must be obtained. Here we focus on developing a probabilistic framework for analysis of uncertainties occurring when satellite data (e.g., MODIS) are employed to retrieve biophysical properties of vegetation. Indeed, the process of remotely estimating vegetation properties involves inverting a Radiative Transfer Model (RTM), as in the case of the MOD15 algorithm where seven atmospherically corrected reflectance factors are ingested and compared to a set of computed, RTM-based, reflectances (look-up table) to infer the Leaf Area Index (LAI). Since inversion is generally ill-conditioned, and since a-priori information is important in constraining the inverse model, sensitivity analysis plays a key role in defining which parameters have the greatest impact to the computed observation. We develop a framework to perform global sensitivity analysis, i.e., to determine how the output changes as all inputs vary continuously. We used a coupled Leaf-Canopy radiative transfer Model (LCM) to approximate the functional relationship between the observed reflectance and vegetation biophysical parameters. LCM was designed to study the feasibility of detecting leaf/canopy biochemistry using remote sensed observations and has the unique capability to include leaf biochemistry (e.g., chlorophyll, water, lignin, protein) as input parameters. The influence of LCM input parameters (including canopy morphological and biochemical parameters) on the hemispherical reflectance is captured by computing the "main effects", which give information about the influence of each input, and the "sensitivity indices", i.e., the expected amount by which the uncertainty in the output is reduced if the true value of a specific input was known. Since RTMs are generally computationally expensive, we develop a Gaussian Process (GP) statistical model to approximate the LCM output surface. Once the GP parameters are estimated (using simulated data from the LCM model), the computation of main effects and sensitivity indices is straightforward. Using this approach, we were able to quantify the importance of LCM input parameters for each of the eight wavelengths centered on the MODIS bands commonly used to observe vegetation (visible and Near-Infrared or NIR). We found that chlorophyll dominates the visible region while LAI has the largest effect on the NIR. Surprisingly, we also found that lignin is sensitive in short-wave infrared (1640nm and 2310nm) where it is the main contributor to the overall reflectance. These results indicate the feasibility and utility of probabilistic uncertainty and sensitivity analysis for RTMs. The developed methodology can be used both to improve RTMs for better measurement predictions, and to guide the data collection or land cover to reduce the level of uncertainty.
B24B-06
Hyperspectral Study of the Arctic Tundra Ecosystem Using an Automated Robotic Cart System
Our study in the NSF Biocomplexity project is carried on by collecting spectral data with the help of an automated robotic tram system over the drying arctic lake bed. The robotic cart samples three 300 meter long transects spread across the lake basin automatically, taking reflectance measurements at each meter using a dual detector spectrometer designed to correct for changing sky conditions. Surface reflectance data were collected for three consecutive years for 2005, 2006 and 2007 as part of the project, which provides a baseline dataset of surface conditions. Three spectral indices, Normalized Difference Vegetation Index (NDVI), a measure of vegetation ‘greenness', the Photochemical Reflectance Index (PRI), a measure of carotenoid pigment levels, and the Water Band Index (WBI), a measure of vegetation moisture content, are calculated from the optical data collected to study the surface conditions of the lakebed. Comparison of three years NDVI data showed different greenness conditions of the surface. Peak season NDVI values were the lowest in 2005 compared to 2006 and 2007. WBI values for the two dry years 2005 and 2007 were similar for all the tramlines except for the beginning of the season. PRI values for the two dry years 2005 and 2007 had similar trends for all the tramlines except for the beginning of the season. This tram system along with the cyberinfrastucture tools that we are developing gives us the opportunity for developing a technology to the next level to facilitate the research in the field of environmental science and terrestrial ecology.
B24B-07
Cropland Area Extraction in China with Multi-Temporal MODIS Data
Abstract: extracting the area of cropland in China is very important for agricultural management, land degradation and ecosystem assessment. In this study we investigate the potential and the methodology for the cropland area extraction using multi-temporal MODIS EVI data and some ancillary data. A 16-day composite EVI time-series data for 2003 (6 March 2003 - 2 December 2003) with a spatial resolution of 500 m, and the ancillary data included Land-use GIS data, Landsat TM/ETM, ASTER data, and county-level cultivated land statistical data of year 2000. The Self-Organizing Map (SOM) neural network classification algorithm was applied to the EVI data set. To focus on agricultural and desertification, we designed 9 land-cover types: 1) water, 2) woodland, 3) grassland, 4) dry cropland, 5) sandy, 6) paddy, 7) wetland, 8) urban/bare, and 9) snow/ice. The overall classification accuracy was 85% with a kappa coefficient of 0.84. The EVI data sets were sensitive and performed well in distinguishing the majority of land cover types. We also used county-level cultivated land statistical data from the year 2000 to evaluate the accuracy of the agricultural area from classification results, and found that the correlation coefficient was high in most counties. The result of this study shows that the methodology used in this study is, in general, feasible for cropland extraction in China. Keywords: MODIS, EVI, SOM, Cropland, land cover.
B24B-08
Employing UAVs to Acquire Detailed Vegetation and Bare Ground Data for Assessing Rangeland Health
Because of its value as a historical record (extending back to the mid 1930s), aerial photography is an important tool used in many rangeland studies. However, these historical photos are not very useful for detailed analysis of rangeland health because of inadequate spatial resolution and scheduling limitations. These issues are now being resolved by using Unmanned Aerial Vehicles (UAVs) over rangeland study areas. Spatial resolution improvements have been rapid in the last 10 years from the QuickBird satellite through improved aerial photography to the new UAV coverage and have utilized improved sensors and the more simplistic approach of low altitude flights. Our rangeland health experiments have shown that the low altitude UAV digital photography is preferred by rangeland scientists because it allows, for the first time, their identification of vegetation and land surface patterns and patches, gap sizes, bare soil percentages, and vegetation type. This hyperspatial imagery (imagery with a resolution finer than the object of interest) is obtained at about 5cm resolution by flying at an altitude of 150m above the surface of the Jornada Experimental Range in southern New Mexico. Additionally, the UAV provides improved temporal flexibility, such as flights immediately following fires, floods, and other catastrophic disturbances, because the flight capability is located near the study area and the vehicles are under the direct control of the users, eliminating the additional steps associated with budgets and contracts. There are significant challenges to improve the data to make them useful for operational agencies, namely, image distortion with inexpensive, consumer grade digital cameras, difficulty in detecting sufficient ground control points in small scenes (152m by 114m), accuracy of exterior UAV information on X,Y, Z, roll, pitch, and heading, the sheer number of images collected, and developing reliable relationships with ground-based data across a broad range of topographies and plant communities. Our efforts are currently focused on developing a complete and efficient workflow for UAV operational missions consisting of flight planning, image acquisition, image rectification and mosaicking, and image classification. The remote sensing capability is being incorporated into existing rangeland health assessment and monitoring protocols.