Mapping Mangrove Condition

Map of Ten Thousand Islands, southwest Florida

The Science Issue and Relevance: Mangroves have decreased worldwide due to human development, climate change and other forces. In southwest Florida, tremendous growth and development pressure has resulted in appreciable losses in mangrove wetlands. Further compounding these human induced stresses, eustatic sea level rise has serious implications for these ecosystems. A technology is urgently needed that can detect adverse change in mangrove condition while mitigation is still possible.

Methodology for Addressing the Issue: Our approach relied on creating remote sensing techniques that could extract all the needed biophysical information from optical reflectance image data. Use of remote sensing offered a convenient method to timely monitor mangroves over large regions, thereby, increasing the likelihood of detection before irreversible loss. Direct assessment of mangrove condition depended on mapping key biophysical indicators: mangrove canopy leaf density (LAI), average orientation (LAD), and leaf reflectance. To accomplish this, a light interaction model was created and tested with field data obtained from 20 red, black, and white mangrove field sites in the Ten Thousand Island region of southwest Florida.

Figure 1. Basin mangrove leaf reflectance. Figure 2. Mangrove leaf reflectance variability.

Analyses of the field data showed little difference between red, black, and white mangrove leaves at each site (Figure 1) and high variability across all 20 sites (Figure 2). Leaf reflectance is the dominant way to detect the early onset of change in vegetation. Tracking leaf reflectance change would provide an early indicator of adverse change in mangrove condition.

Simulation of satellite remote sensing was achieved with light detection instruments onboard a helicopter platform above the mangrove canopy. The range in obtained canopy reflectance (Figure 3) portrays the high spectral variability of mangroves found in basin, riverine, and overwash settings. Using the simulated satellite canopy reflectance as input, the light interaction model calculated sets of LAI, LAD, and leaf reflectance spectra while predicting the canopy reflectance spectra with an accuracy of 97% as illustrated in Figure 4.

Figure 3. Observed mangrove canopy reflectance. Figure 4. Predicted mangrove canopy reflectance.

After leaf reflectance, the most effective indicator of mangrove condition determinable from optical remote sensing data is change in LAI and LAD. The high performance of light interaction model performance was again demonstrated by its high 93% prediction accuracy of LAI’s and LAD as compared to field observed values (Figure 5). In addition, the predicted central tendency of LAD was spherical as expected. Predicted LAI’s were also highly aligned with a vegetation index (NDVI) calculated from broadband satellite data (Figure 6).

Future Steps: Leaf reflectance changes throughout the 20 sites may reflect natural variability; however, predicted leaf reflectance could reveal a departure from healthy condition providing an early indicator of mangrove degradation. Further research would substantiate that capability.

Figure 5. Modelled mangrove canopy LAI. Figure 6. Satellite data predicted mangrove LAI.

Keywords: Remote Sensing, Mangrove Condition, Leaf Reflectance, Leaf Area Index, LAI 

References

Ramsey III, E., and A. Rangoonwala, 2011, Remote Sensing of Wetland Vegetation Focusing on Hyperspectral Mapping. P. S. Thenkabail, ed, J. G. Lyon and Prof. A. Huete, coeds. Hyperspectral Remote Sensing of Vegetation, CRC Press, pp. 487–511.

Ramsey III, E., and Jensen, J., 1996, Remote sensing of mangroves: relating canopy spectra to site-specific data, Photogramm Eng Rem S, 62(8):939-948.

Ramsey III, E., and Jensen, J., 1995, Modeling mangrove canopy reflectance using a light interaction model and an optimization technique. J. Lyon and J. McCarthy, eds. Wetland and Environmental Applications of GIS. CRC Press, Inc., Boca Raton, Florida, pp. 61-81.