SPIE Membership Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:
SPIE Photonics West 2019 | Call for Papers

2018 SPIE Optics + Photonics | Register Today



Print PageEmail PageView PDF

Remote Sensing

Moderate resolution satellite data for mapping salt marshes

Accurate in situ information plays a crucial role in calibrating models to estimate and map salt marsh characteristics using satellite data.
16 April 2013, SPIE Newsroom. DOI: 10.1117/2.1201304.004800

Biophysical characteristics of salt marshes are primary indicators of their photosynthetic capacity, nitrogen content, and physiological status. These include the four main characteristics of canopy chlorophyll content: green leaf area index (ratio of green foliage area to ground area), green vegetation fraction (percent green canopy cover), green biomass (which also act as proxies for estimating net primary productivity —NPP), and carbon sequestration potential. Monitoring these characteristics through remotely sensed data can help infer the overall condition and productivity of these valuable natural resources on a larger scale. This would allow effective management strategies in high priority areas. Our research aims to enhance the marsh monitoring practices by including biophysical characteristics (the true early indicators of marsh health status) derived accurately from information such as NASA's moderate resolution imaging spectroradiometer (MODIS) 250m and 500m data.

Our objectives were to calibrate and test a suite of MODIS-based vegetation indices (VIs) and develop prototype weekly composites of the salt marsh biophysical characteristics for the entire US Gulf Coast from 2000 to 2010. We then planned to perform a time-series analysis to study the overall trend of salt marsh productivity during the last decade. These VIs (normal difference, wide dynamic range, enhanced, soil-adjusted, etc.) have been widely used and tested for monitoring terrestrial vegetation, but not for salt marsh ecosystems. Our VI calibration and tuning came from establishing relationships between MODIS 250m or 500m data and field data in order to predict the four marsh biophysical characteristics mentioned above.

Figure 1. (a) The dual-headed OceanOptics sensor mounted on a 5m (16ft) high frame. (b, c) Instantaneous field of view (spatial resolution) of the sensor (diameter: 2.2m). (d) In situ sensor calibration.

We used atmospherically corrected MODIS 8-day surface reflectance products (250m and 500m), and in situ measurements of biophysical data from 182 locations covering four states (LA, MS, AL, FL). Preliminary analysis revealed that the accuracy of the MODIS-based salt marsh VI calibration depends primarily on three factors: the accuracy of the field data collection, the number of field samples within a MODIS pixel (for accurate representation), and the nature of the MODIS pixel in terms of homogeneity of marsh species, growth status, and absence of large open water areas. We conducted 14 field campaigns covering two dominant salt marsh species (Spartina alterniflora, Juncus romerianus) in the region. The in situ data collected (see Figure 1) from each location included:1 top of canopy reflectance using a hyperspectral radiometer; leaf-level chlorophyll content reading using a chlorophyll meter;2 green leaf area index readings using a leaf area index plant canopy analyzer,3 ceptometer,4,5 and digital photographs; canopy-level chlorophyll content using chlorophyllupper× leaf area index techniques (the technique varies from species to species);4 vegetation fraction readings using digital photographs;6 and above-ground biomass (g/m2). A detailed explanation of the field data collection techniques can be found in a previously published study where we used similar techniques to model the impact of the British Petroleum oil spill in the Gulf of Mexico on salt marshes in Louisiana.7

Based on comparative correlation coefficients (r2), the percent root mean squared error (%RMSE), and the residual trend's wide dynamic range VI, we chose (α=0.1, a red and near IR band normalization index with a fractional coefficient applied to the near IR band)8 to estimate biophysical parameters from the 250m dataset. We chose a visible atmospheric resistant index9 to predict biophysical parameters from the 500m dataset. After successful calibration and validation, we used the best-fit models to develop time series composites of MODIS-based biophysical map products for the growing seasons (April-October) from 2000–2010, both for the 250m and 500m datasets (see Figure 2). Approximately 2500 maps were generated combining 250m and 500m resolution data for the 10-year duration. These maps enable the end user to analyze the temporal variations in biophysical characteristics observed in the salt marsh habitat in the northern Gulf of Mexico during the last decade. The maps also highlight areas that show signs of degradation after a natural or anthropogenic event, indicating that they are in need of immediate restoration or that the effectiveness of the previous restoration projects needs to be monitored. We constructed phenological plots for selected locations in order to study biophysical characteristic variations over the growing seasons from 2000–2010 (see Figure 3).

Figure 2. Composites of biophysical characteristics using NASA's moderate resolution imaging spectroradiometer or MODIS 250m (a,c,e,g) and 500m (b,d,f,h) data.

Figure 3. (a-d): Phenological variations in Plaquemines Parish, LA, salt marshes from 2000–2011 (250m MODIS data). Similar trends were observed in the 500m MODIS data.

Figure 4. (a-b): Installation of an open path CO2/H2O analyzer.

Our ultimate goal is to use these biophysical characteristics in NPP prediction models to produce maps of spatial distribution (species dependent and species invariant) and analyze the changes in carbon sequestration potential of the salt marshes after natural and anthropogenic disasters. Therefore, we have been conducting experiments at the same sampling locations used for MODIS calibrations to acquire atmospheric and soil CO2 flux data using eddy covariance and soil flux instruments. In natural conditions, photosynthesis depends on CO2coming from three sources, the atmosphere, soil, and the plant itself. Assuming that CO2 released by respiration is merely reused in photosynthesis, we intend to implement a combination automated soil flux system and open-path CO2/H2O analyzer10–12 to estimate soil respiration and atmospheric CO2 flux. The instruments will provide reasonable estimates of the net ecosystem exchange of CO2 between the soil/atmosphere and vegetation/atmosphere interface (see Figure 4). Our future research will combine this net ecosystem exchange flux data with marsh biophysical data to map the spatio-temporal variability of NPP and carbon sequestration potential in the salt marsh ecosystem.

Deepak Mishra, Shuvankar Ghosh
Department of Geography
University of Georgia (UGA)
Athens, GA

Deepak Mishra, assistant professor, department of geography, directs the environmental remote sensing and spectroscopy labs at UGA. He holds a PhD in natural resources from the University of Nebraska, Lincoln.

Shuvnakar Ghosh is a PhD student in geography whose dissertation research is focused on remote sensing of salt marshes.

1. D. R. Mishra, H. J. Cho, S. Ghosh, C. Downs, A. Fox, P. B. T. Merani, P. Kirui, Post-spill state of the marsh: Remote estimation of the ecological impact of the Gulf of Mexico oil spill on Louisiana salt marshes, Remote Sensing of Environ. 118, p. 176-185, 2012. doi:10.1016/j.rse.2011.11.007
2. A. A. Gitelson, A. Vina, V. Ciganda, D. C. Rundquist, Remote estimation of canopy chlorophyll content in crops, Geophys. Res. Lett. 32, p. 1-4, 2005.
3. S. Malone, D. A. Herbert, D. L. Holshouser, Evaluation of the LAI-2000 plant canopy analyzer to estimate leaf area in manually defoliated soybean, Agronomy J. 94, p. 1012-1019, 2002.
4. S. Delalieux, B. Somers, S. Hereijgers, W. W. Verstraeten, W. Keulemans, P. Coppi, A near-IR narrow waveband ratio to determine leaf area index in orchards, Remote Sensing of Environ. 112, p. 3762-3772, 2008.
5. J. M. Kovacs, J. M. L. King, F. F. de Santiago, F. Flores-Verdugo, Evaluating the condition of a mangrove forest of the Mexican Pacific based on an estimated leaf area index mapping approach, Environ. Monitoring and Assess. 157, p. 137-49, 2009.
6. A. A. Gitelson, Y. J. Kaufman, R. Stark, D. Rundquist, Novel algorithms for remote estimation of vegetation fraction, Remote Sensing of Environ. 80, p. 76-87, 2002.
7. D. R. Mishra, H. J. Cho, S. Ghosh, A. A. Fox, C. Downs, P. B. T. Merani, P. Kirui, Post-spill state of the marsh: Impact of the Gulf of Mexico oil spill on the health and productivity of Louisiana salt marshes, Remote Sensing of Environ. 118, p. 176-185, 2012.
8. A. A. Gitelson, Wide dynamic range vegetation index for remote quantification of crop biophysical characteristics, J. Plant Physiol. 161, p. 165-173, 2004.
9. A. Vina, G. Henebry, A. A. Gitelson, Satellite monitoring of vegetation dynamics: sensitivity enhancement by the wide dynamic range vegetation index, Geophys. Res. Lett. 31, p. L04503, 2004.
10. S. Rutledge, D. I. Campbell, Dennis Baldocchi, L. A. Schipper, Photodegradation leads to increased carbon dioxide losses from terrestrial organic matter, Global Change Biol. 16, p. 3065-3074, 2010.
11. R. Madsen, L. Xu, B. Claassen, D. McDermitt, Surface monitoring method for carbon capture and storage projects, Energy Procedia 1, p. 2161-2168, 2009.
12. J. A. Bunce, Performance characteristics of an area distributed free air carbon dioxide enrichment system, Agricultural and Forest Meteorology 151, p. 1152-1157, 2011.