LAI inversion algorithm based on directional reflectance kernels

Leaf area index (LAI) is an important ecological and environmental parameter. A new LAI algorithm is developed using the principles of ground LAI measurements based on canopy gap fraction. First, the relationship between LAI and gap fraction at various zenith angles is derived from the definition of LAI. Then, the directional gap fraction is acquired from a remote sensing bidirectional reflectance distribution function (BRDF) product. This acquisition is obtained by using a kernel driven model and a large-scale directional gap fraction algorithm. The algorithm has been applied to estimate a LAI distribution in China in mid-July 2002. The ground data acquired from two field experiments in Changbai Mountain and Qilian Mountain were used to validate the algorithm. To resolve the scale discrepancy between high resolution ground observations and low resolution remote sensing data, two TM images with a resolution approaching the size of ground plots were used to relate the coarse resolution LAI map to ground measurements. First, an empirical relationship between the measured LAI and a vegetation index was established. Next, a high resolution LAI map was generated using the relationship. The LAI value of a low resolution pixel was calculated from the area-weighted sum of high resolution LAIs composing the low resolution pixel. The results of this comparison showed that the inversion algorithm has an accuracy of 82%. Factors that may influence the accuracy are also discussed in this paper.     

Measurements and simulation of forest leaf area index and net primary productivity in Northern China

Large scale process-based modeling is a useful approach to estimate distributions of global net primary productivity (NPP). In this paper, in order to validate an existing NPP model with observed data at site level, field experiments were conducted at three sites in northern China. One site is located in Qilian Mountain in Gansu Province, and the other two sites are in Changbaishan Natural Reserve and Dunhua County in Jilin Province. Detailed field experiments are discussed and field data are used to validate the simulated NPP. Remotely sensed images including Landsat Enhanced Thematic Mapper plus (ETM+, 30 m spatial resolution in visible and near infrared bands) and Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER, 15m spatial resolution in visible and near infrared bands) are used to derive maps of land cover, leaf area index, and biomass. Based on these maps, field measured data, soil texture and daily meteorological data, NPP of these sites are simulated for year 2001 with the boreal ecosystem productivity simulator (BEPS). The NPP in these sites ranges from 80 to 800 gCm(-2)a(-1). The observed NPP agrees well with the modeled NPP. This study suggests that BEPS can be used to estimate NPP in northern China if remotely sensed images of high spatial resolution are available.     

Application of a new leaf area index algorithm to China's landmass using MODIS data for carbon cycle research

An operational system was developed for mapping the leaf area index (LAI) for carbon cycle models from the moderate resolution imaging spectroradiometer (MODIS) data. The LAI retrieval algorithm is based on Deng et al. [2006. Algorithm for global leaf area index retrieval using satellite imagery. IEEE Transactions on Geoscience and Remote Sensing, 44, 2219-2229], which uses the 4-scale radiative transfer model [Chen, J.M., Leblancs, 1997. A 4-scale bidirectional reflection model based on canopy architecture. IEEE Transactions on Geoscience and Remote Sensing, 35, 1316-1337] to simulate the relationship of LAI with vegetated surface reflectance measured from space for various spectral bands and solar and view angles. This algorithm has been integrated to the MODISoft platform, a software system designed for processing MODIS data, to generate 250 m, 500 m and 1 km resolution LAI products covering all of China from MODIS MOD02 or MOD09 products. The multi-temporal interpolation method was implemented to remove the residual cloud and other noise in the final LAI product so that it can be directly used in carbon models without further processing. The retrieval uncertainties from land cover data were evaluated using five different data sets available in China. The results showed that mean LAI discrepancies can reach 27%. The current product was also compared with the NASA MODIS MOD15 LAI product to determine the agreement and disagreement of two different product series. LAI values in the MODIS product were found to be 21% larger than those in the new product. These LAI products were compared against ground TRAC measurements in forests in Qilian Mountain and Changbaishan. On average, the new LAI product agrees with the field measurement in Changbaishan within 2%, but the MODIS product is positively biased by about 20%. In Qilian Mountain, where forests are sparse, the new product is lower than field measurements by about 38%, while the MODIS product is larger by about 65%.     

A Comparison of Spatial and Spectral Image Resolution for Mapping Invasive Plants in Coastal California

We explored the potential of detecting three target invasive species: iceplant (Carpobrotus edulis), jubata grass (Cortaderia jubata), and blue gum (Eucalyptus globulus) at Vandenberg Air Force Base, California. We compared the accuracy of mapping six communities (intact coastal scrub, iceplant invaded coastal scrub, iceplant invaded chaparral, jubata grass invaded chaparral, blue gum invaded chaparral, and intact chaparral) using four images with different combinations of spatial and spectral resolution: hyperspectral AVIRIS imagery (174 wavebands, 4 m spatial resolution), spatially degraded AVIRIS (174 bands, 30 m), spectrally degraded AVIRIS (6 bands, 4 m), and both spatially and spectrally degraded AVIRIS (6 bands, 30 m, i.e., simulated Landsat ETM data). Overall success rates for classifying the six classes was 75% (kappa 0.7) using full resolution AVIRIS, 58% (kappa 0.5) for the spatially degraded AVIRIS, 42% (kappa 0.3) for the spectrally degraded AVIRIS, and 37% (kappa 0.3) for the spatially and spectrally degraded AVIRIS. A true Landsat ETM image was also classified to illustrate that the results from the simulated ETM data were representative, which provided an accuracy of 50% (kappa 0.4). Mapping accuracies using different resolution images are evaluated in the context of community heterogeneity (species richness, diversity, and percent species cover). Findings illustrate that higher mapping accuracies are achieved with images possessing high spectral resolution, thus capturing information across the visible and reflected infrared solar spectrum. Understanding the tradeoffs in spectral and spatial resolution can assist land managers in deciding the most appropriate imagery with respect to target invasives and community characteristics.     

Retrieving leaf area index for coniferous forest in Xingguo County, China with Landsat ETM+ images

Spatial distributions of the leaf area index (LAI) needed for carbon cycle modeling in Xingguo County, China were estimated based on correlations between the field-measurements and vegetation indices (VIs). After making geometric and atmospheric corrections to two Landsat ETM+ images, one in January 2000 and the other in May 2003, three VIs (SR, NDVI, and RSR) were derived, and their separate correlations with ground LAI measurements were established. The correlation with RSR was the highest among the three VIs. The retrieved LAI values for January 2000 were lower than those for May 2003 because of a small seasonal variation in the coniferous forests (predominantly masson pine) and the decrease in the understorey vegetation during winter.     

Observation and simulation of net primary productivity in Qilian Mountain, western China

We modeled net primary productivity (NPP) at high spatial resolution using an advanced spaceborne thermal emission and reflection radiometer (ASTER) image of a Qilian Mountain study area using the boreal ecosystem productivity simulator (BEPS). Two key driving variables of the model, leaf area index (LAI) and land cover type, were derived from ASTER and moderate resolution imaging spectroradiometer (MODIS) data. Other spatially explicit inputs included daily meteorological data (radiation, precipitation, temperature, humidity), available soil water holding capacity (AWC), and forest biomass. NPP was estimated for coniferous forests and other land cover types in the study area. The result showed that NPP of coniferous forests in the study area was about 4.4 tCha(-1)y(-1). The correlation coefficient between the modeled NPP and ground measurements was 0.84, with a mean relative error of about 13.9%.     

Landsat ETM+ Images in the Estimation of Seasonal Lake Water Quality in Boreal River Basins

We investigated the use of Landsat ETM+ images in the monitoring of turbidity, colored dissolved organic matter (CDOM), and Secchi disk transparency (Z(SD)) in lakes of two river basins located in southern Finland. The ETM+ images were acquired in May, June, and September 2002 and were corrected for atmospheric disturbance using the simplified method of atmospheric correction (SMAC) model. The in situ measurements consisted of water sampling in the largest lake of the region, routine monitoring results for the whole study area, and Z(SD) observations made by volunteers. The ranges of the water quality variables in the dataset were as follows: turbidity, 0.6-25 FNU; absorption coefficient of CDOM at 400 nm, 1.0-12.2 m(-1); Z(SD), 0.5-5.5 m; and chlorophyll a concentration, 2.4-80 mug L(-1). The estimation accuracies of the image-specific empirical algorithms expressed as relative errors were 23.0% for turbidity, 17.4% for CDOM, and 21.1% for Z(SD). If concurrent in situ measurements had not been used for algorithm training, the average error would have been about 37%. The atmospheric correction improved the estimation accuracy only slightly compared with the use of top-of-atmospheric reflectances. The accuracy of the water quality estimates without concurrent in situ measurements could have been improved if in-image atmospheric parameters had been available. The underwater reflectance simulations of the ETM+ channel wavelengths using water quality typical for Finnish lakes (data from 1113 lakes) indicated that region-specific algorithms may be needed in other parts of the country, particularly in the case of Z(SD). Despite the limitations in the spectral and radiometric resolutions, ETM+ imagery can be an effective aid, particularly in the monitoring and management of small lakes (<1 km(2)), which are often not included in routine monitoring programs.     

Combining remote sensing imagery and forest age inventory for biomass mapping

Aboveground biomass (AGB) of forests is an important component of the global carbon cycle. In this study, Landsat ETM(+) images and field forest inventory data were used to estimate AGB of forests in Liping County, Guizhou Province, China. Three different vegetation indices, including simple ratio (SR), reduced simple ratio (RSR), and normalized difference vegetation index (NDVI), were calculated from atmospherically corrected ETM(+) reflectance images. A leaf area index (LAI) map was produced from the RSR map using a regression model based on measured LAI and RSR. The LAI map was then used to develop an initial AGB map, from which forest stand age was deduced. Vegetation indices, LAI, and forest stand age were together used to develop AGB estimation models for different forest types through a stepwise regression analysis. Significant predictors of AGB changed with forest types. LAI and NDVI were significant predictors of AGB for Chinese fir (R(2)=0.93). The model using LAI and stand age as predictors explained 94% of the AGB variance for coniferous forests. Stand age captured 79% of the AGB variance for broadleaved forests (R(2)=0.792). AGB of mixed forests was predicted well by LAI and SR (R(2)=0.931). Without differentiating among forest types, the model with SR and LAI as predictors was able to explain 90% of AGB variances of all forests. In Liping County, AGB shows a strong gradient that increases from northeast to southwest. About 64% of the forests have AGB in the range from 90 to 180 t ha(-1).     

WATERSHED WEIGHTING OF EXPORT COEFFICIENTS TO MAP CRITICAL PHOSPHOROUS LOADING AREAS1

ABSTRACT: The Export Coefficient model (ECM) is capable of generating reasonable estimates of annual phosphorous loading simply from a watershed's land cover data and export coefficient values (ECVs). In its current form, the ECM assumes that ECVs are homogeneous within each land cover type, yet basic nutrient runoff and hydrological theory suggests that runoff rates have spatial patterns controlled by loading and filtering along the flow paths from the upslope contributing area and downslope dispersal area. Using a geographic information system (GIS) raster, or pixel, modeling format, these contributing area and dispersal area (CADA) controls were derived from the perspective of each individual watershed pixel to weight the otherwise homogeneous ECVs for phosphorous. Although the CADA‐ECM predicts export coefficient spatial variation for a single land use type, the lumped basin load is unaffected by weighting. After CADA weighting, a map of the new ECVs addressed the three fundamental criteria for targeting critical pollutant loading areas: (1) the presence of the pollutant, (2) the likelihood for runoff to carry the pollutant offsite, and (3) the likelihood that buffers will trap nutrients prior to their runoff into the receiving water body. These spatially distributed maps of the most important pollutant management areas were used within New York's West Branch Delaware River watershed to demonstrate how the CADA‐ECM could be applied in targeting phosphorous critical loading areas.     

DEM AGGREGATION FOR WATERSHED MODELING1

ABSTRACT: The widely available USGS 7.5‐minute Digital Elevation Model (DEM) has a cell size of approximately 30 m × 30 m. This high resolution topographic information is impractical for many applications of distributed hydrologic and water quality models. In this study, cells were aggregated into coarse‐resolution areal units, termed grids, and a method to approximate flow direction for coarse‐resolution grids from 30 m DEM cells was developed. The method considers the flow path defined from the fine‐resolution DEM in determining a grid's flow direction and makes flow directions for grids closely follow the flow pattern suggested by the DEM. The aggregation method was applied to a DEM of Goodwater Creek, a nearly flat watershed that is located in central Missouri. The drainage networks derived for different levels of cell aggregations showed that grid aggregates of the Goodwater Creek watershed provided an adequate representation of the landscape topography.     

Annual Estimates of Recharge,Quick‐Flow Runoff,and Evapotranspiration for the Contiguous U.S. Using Empirical Regression Equations

This study presents new data‐driven, annual estimates of the division of precipitation into the recharge, quick‐flow runoff, and evapotranspiration (ET) water budget components for 2000‐2013 for the contiguous United States (CONUS). The algorithms used to produce these maps ensure water budget consistency over this broad spatial scale, with contributions from precipitation influx attributed to each component at 800 m resolution. The quick‐flow runoff estimates for the contribution to the rapidly varying portion of the hydrograph are produced using data from 1,434 gaged watersheds, and depend on precipitation, soil saturated hydraulic conductivity, and surficial geology type. Evapotranspiration estimates are produced from a regression using water balance data from 679 gaged watersheds and depend on land cover, temperature, and precipitation. The quick‐flow and ET estimates are combined to calculate recharge as the remainder of precipitation. The ET and recharge estimates are checked against independent field data, and the results show good agreement. Comparisons of recharge estimates with groundwater extraction data show that in 15% of the country, groundwater is being extracted at rates higher than the local recharge. These maps of the internally consistent water budget components of recharge, quick‐flow runoff, and ET, being derived from and tested against data, are expected to provide reliable first‐order estimates of these quantities across the CONUS, even where field measurements are sparse.     

Estimating Floodwater Depths from Flood Inundation Maps and Topography

Information on flood inundation extent is important for understanding societal exposure, water storage volumes, flood wave attenuation, future flood hazard, and other variables. A number of organizations now provide flood inundation maps based on satellite remote sensing. These data products can efficiently and accurately provide the areal extent of a flood event, but do not provide floodwater depth, an important attribute for first responders and damage assessment. Here we present a new methodology and a GIS‐based tool, the Floodwater Depth Estimation Tool (FwDET), for estimating floodwater depth based solely on an inundation map and a digital elevation model (DEM). We compare the FwDET results against water depth maps derived from hydraulic simulation of two flood events, a large‐scale event for which we use medium resolution input layer (10 m) and a small‐scale event for which we use a high‐resolution (LiDAR; 1 m) input. Further testing is performed for two inundation maps with a number of challenging features that include a narrow valley, a large reservoir, and an urban setting. The results show FwDET can accurately calculate floodwater depth for diverse flooding scenarios but also leads to considerable bias in locations where the inundation extent does not align well with the DEM. In these locations, manual adjustment or higher spatial resolution input is required.     

MONITORING SUSPENDED PARTICULATE MATTER IN PUERTO RICO: FIELD MEASUREMENTS AND REMOTE SENSING1

ABSTRACT: The spatial distribution of suspended particulatematter (SPM) was estimated in Mayagüez Bay on the west coast of Puerto Rico by using traditional ship board measurements and remotely sensed data acquired over four days during January 1990. This effort was part of a joint project between NASA and the University of Puerto Rico to develop techniques to monitor the water quality of a Caribbean coastal zone. This paper presents the methods and algorithms developed to map and analyze short term changes in the source and spatial distribution of SPM in Mayagüez Bay by using remotely sensed data acquired by the Calibrated Airborne Multispectral Scanner (CAMS). A PC-based data acquisition system was developed to record continuous ship measurements of select in- water variables. Spectral reflectances derived from CAMS red and near-IR data were corrected for atmospheric effects and then used to generate maps of SPM. These maps displayed SPM plumes associated with the mouths of the bay's three river systems. Significant day-to-day differences in the spatial characteristics were observed, suggesting that changes in river discharge occurred. However, an analysis of estimated river discharge, sediment yield, local wind field, and thermal river plume indicates that observed sediment plumes result primarily from wind-driven resuspension events.     

Optimal spatial resolution for collection of ground data and multi-sensor image mapping of a soil erosion cover factor

Military training activities disturb ground and vegetation cover of landscapes and increases potential soil erosion. To monitor the dynamics of soil erosion, there is an important need for an optimal sampling design in which determining the optimal spatial resolutions in terms of size of sample plots used for the collection of ground data and the size of pixels for mapping. Given a sample size, an optimal spatial resolution should be cost-efficient in both sampling costs and map accuracy. This study presents a spatial variability-based method for that purpose and compared it with the traditional methods in a study area in which a soil erosion cover factor was sampled and mapped with multiple plot sizes and multi-sensor images. The results showed that the optimal spatial resolutions obtained using the spatial variability-based method were 12 and 20m for years 1999 and 2000, respectively, and were consistent with those using the traditional methods. Moreover, the most appropriate spatial resolutions using the high-resolution images were also consistent with those using ground sample data, which provides a potential to use the high-resolution images instead of ground data to determine the optimal spatial resolutions before sampling. The most appropriate spatial resolutions above were then verified in terms of cost-efficiency which was defined as the product of sampling cost and map error using ordinary kriging without images and sequential Gaussian co-simulation with images to generate maps.     

Evaluation of Methods for Delineating Riparian Zones in a Semi‐Arid Montane Watershed

Riparian zones in semi‐arid, mountainous regions provide a disproportionate amount of the available wildlife habitat and ecosystem services. Despite their importance, there is little guidance on the best way to map riparian zones for broad spatial extents (e.g., large watersheds) when detailed maps from field data or high‐resolution imagery and terrain data are not available. Using well‐established accuracy metrics (e.g., kappa, precision, computational complexity), we evaluated eight methods commonly used to map riparian zones. Focusing on a semi‐arid, mountainous watershed, we found that the most accurate and robust method for mapping riparian zones combines data on upstream drainage area and valley topography. That method performed best regardless of stream order, and was most effective when implemented with fine resolution topographic and stream line data. Other commonly used methods to model riparian zones, such as those based on fixed‐width buffers, yielded inaccurate results. We recommend that until very‐high resolution (<1 m) elevation data are available at broad extents, models of riparian zones for semi‐arid mountainous regions should incorporate drainage area, valley topography, and quantify uncertainty.     

Modeling Streams and Hydrogeomorphic Attributes in Oregon From Digital and Field Data1

Abstract: Managers, regulators, and researchers of aquatic ecosystems are increasingly pressed to consider large areas. However, accurate stream maps with geo‐referenced attributes are uncommon over relevant spatial extents. Field inventories provide high‐quality data, particularly for habitat characteristics at fine spatial resolutions (e.g., large wood), but are costly and so cover relatively small areas. Recent availability of regional digital data and Geographic Information Systems software has advanced capabilities to delineate stream networks and estimate coarse‐resolution hydrogeomorphic attributes (e.g., gradient). A spatially comprehensive coverage results, but types of modeled outputs may be limited and their accuracy is typically unknown. Capitalizing on strengths in both field and regional digital data, we modeled a synthetic stream network and a variety of hydrogeomorphic attributes for the Oregon Coastal Province. The synthetic network, encompassing 96,000 km of stream, was derived from digital elevation data. We used high‐resolution but spatially restricted data from field inventories and streamflow gauges to evaluate, calibrate, and interpret hydrogeomorphic attributes modeled from digital elevation and precipitation data. The attributes we chose to model (drainage area, mean annual precipitation, mean annual flow, probability of perennial flow, channel gradient, active‐channel width and depth, valley‐floor width, valley‐width index, and valley constraint) have demonstrated value for stream research and management. For most of these attributes, field‐measured, and modeled values were highly correlated, yielding confidence in the modeled outputs. The modeled stream network and attributes have been used for a variety of purposes, including mapping riparian areas, identifying headwater streams likely to transport debris flows, and characterizing the potential of streams to provide high‐quality habitat for salmonids. Our framework and models can be adapted and applied to areas where the necessary field and digital data exist or can be obtained.     

Effect of the Spatial Variability of Land Use,Soil Type,and Precipitation on Streamflows in Small Watersheds1

Abstract: The spatial variability of the data used in models includes the spatial discretization of the system into subsystems, the data resolution, and the spatial distribution of hydrologic features and parameters. In this study, we investigate the effect of the spatial distribution of land use, soil type, and precipitation on the simulated flows at the outlet of “small watersheds” (i.e., watersheds with times of concentration shorter than the model computational time step). The Soil and Water Assessment Tool model was used to estimate runoff and hydrographs. Different representations of the spatial data resulted in comparable model performances and even the use of uniform land use and soil type maps, instead of spatially distributed, was not noticeable. It was found that, although spatially distributed data help understand the characteristics of the watershed and provide valuable information to distributed hydrologic models, when the watershed is small, realistic representations of the spatial data do not necessarily improve the model performance. The results obtained from this study provide insights on the relevance of taking into account the spatial distribution of land use, soil type, and precipitation when modeling small watersheds.     

STREAM HEALTH RANKINGS PREDICTED BY SATELLITE DERIVED LAND COVER METRICS1

ABSTRACT: Land cover and land use change have long been known to influence the chemical, physical, and biological characteristics of streams. This study makes use of land cover maps derived from fine resolution satellite imagery and an extensive stream quality dataset to determine the relationship between small watershed health rankings and land cover composition and configuration. Landscape metrics were derived from digital impervious surface area (ISA), tree cover (percent), and agricultural crop maps within Montgomery County, Maryland. Watershed rankings were developed by state and county collaborators (MD‐DNR and MCDEP) using extensive biological and chemical measurements. In stepwise logistic regression models the factors accounting for the most variation in stream health ranking were the percent ISA, followed by the percent of tree cover. Riparian buffer zone tree cover was also a significant predictor. Of the metrics that considered the spatial configuration of the landscape, a contagion index and the percent of ISA in the flow path from the ISA to the stream were also found to be significant predictors of stream health. Despite limited ability to characterize landscape configuration or narrow riparian buffer zone vegetation with coarser resolution imagery (from Landsat), model results were not significantly different from those based on the use of fine‐resolution ISA information, suggesting that broader area applications of the approach are possible. The results indicate that management practices designed to improve stream water quality should focus on the amount of ISA and tree cover in both the watershed and within the buffer zone.     

Interpolating SRTM Elevation Data to Higher Resolution to Improve Hydrologic Analysis

The Shuttle Radar Topography Mission (SRTM) digital elevation model (DEM) has been a valuable resource for hydrological analysis, providing elevation data at a consistent resolution on a near‐global scale. However, its resolution (three arc‐second or 90 m) is sometimes too low to obtain the desired level of accuracy and precision for hydrologic analysis. We evaluated the performance of several methods for interpolating SRTM three arc‐second data to a 30‐m resolution grid to better represent topography and derive terrain characteristics of the landscape. STRM data were interpolated to 30‐m DEMs on a common grid using spline, inverse distance weighting (IDW), kriging (KR), natural neighbor methods, and cubic convolution (CC) resampling. Accuracy of the methods was assessed by comparing interpolated and resampled 30‐m grids with the reference data. Slope, aspect, sinks, and stream networks were derived for the 30‐m grids and compared on a cell‐by‐cell basis to evaluate their performance in reproducing the derivatives. The comparisons identify spline and KR as the most accurate interpolation methods, of which spline is preferred because of its relative simplicity. IDW provided the greatest bias in all methods with artifacts evident in slope and aspect maps. The performance of CC projection directly to a 30‐m resolution was comparable to spline interpolation, thus is recommended as the most convenient method for interpolating SRTM to a higher resolution.     

DEVELOPMENT OF LARGE SCALE GRIDDED RIVER NETWORKS FROM VECTOR STREAM DATA1

ABSTRACT: The Network Tracing Method (NTM) has been developed to determine gridded coarse river networks for modeling large hydrologic systems. For a coarse resolution grid, the NTM determines the downstream cell of each cell and the distance along the actual meandering flow paths between them. Unlike previously developed methods, the NTM uses fine resolution vector river networks as the source of information of the flow patterns rather than digital elevation models. The main advantage of using vector river networks as input is that they capture the hydrologic terrain features better than topographic data do, particularly in areas of low topographic relief. The NTM was applied to South America with a grid resolution of 1 degree by 1 degree and to the globe with a resolution of 2.815 degrees by 2.8125 degrees. Overall, the method captured the flow patterns well. Generated digital river networks and drainage divides showed minor disagreement with those obtained from existing maps, and most of them were consistent with the resolution of the coarse river network. The majority of estimated basin areas were also close to documented values. River lengths calculated with the NTM, however, were consistently underpredicted.