Home Articles Mapping Cyclone Damage to Mangrove Habitats: An Example from South Florida

Mapping Cyclone Damage to Mangrove Habitats: An Example from South Florida


Douglas O. Fuller
Department of Geography and Regional Studies,
University of Miami
1000 Memorial Drive,
Coral Gables,
FL 33124-2221, USA
[email protected]

The impacts of two hurricanes (Katrina and Wilma) were assessed on 44,390 ha of protected mangroves in southwest Florida using a series of 20m multispectral SPOT and 1-km MODIS images. Established empirical relationships between mangrove leaf area index (LAI) and the normalized difference vegetation index (NDVI) were used to generate four different LAI maps before and after the hurricanes. These maps were compared to gridded LAI data (MOD 15A2) derived from the Moderate Resolution Imaging Spectrometer (MODIS) on board the Terra satellite. A semi-empirical approach based on the GeoSAIL radiative transfer model was also developed to estimate both pre- and post-Hurricane LAI based on SPOT band 2 (red). The results indicated modest agreement (r = 0.53) between one empirical formula and MODIS LAI pre-hurricane; while the GeoSAIL approach produced good agreement between post-hurricane MODIS LAI and SPOT LAI (r = 0.69) among the different methods employed. The results suggest that in the absence of detailed field data a relatively simple radiative transfer model linked to multispectral satellite imagery may provide an effective way to monitor the damage and subsequent recovery of mangrove ecosystems following major disturbances such as tropical cyclones.

Background and Introduction
Mangrove and adjacent wetland ecosystems are among the most productive ecosystems on earth (Odum & McIvor, 1990) but are also among the most vulnerable to anthropogenic activities such as dredge and fill, aquaculture, construction, and other environmental modifications in tropical coastal zones (Hogarth, 1999). Mangrove ecosystems also provide many important functions that enhance overall productivity of estuarine and coastal fisheries; they stabilize and may protect coastlines from erosion, and reduce storm and wave impacts in coastal regions. In addition to direct threats posed by coastal development, future atmospheric warming trends coupled with sea-level rise are likely to exacerbate existing threats caused by local anthropogenic disturbance (Hogarth, 1999).

Biogeographic studies reveal that mangroves canopy height is inversely related to latitude, with taller trees possessing higher biomass in equatorial locations relative to stands found closer to the subtropics (Saenger & Snedaker, 1993). Among the plausible reasons advanced for this geographic trend are land-falling tropical cyclones (Odum & McIvor, 1990), which cause widespread mortality of mangrove trees (Krauss et al., 2005; Piou et al., 2006). In the western hemisphere, major tropical cyclones (or hurricanes) tend to occur most frequently in the northern Caribbean, Gulf of Mexico, and South Florida (Muller & Stone, 2001). In 2004 and 2005 this region experienced an anomalously high number of tropical cyclones, which may be explained by positive warming trends in sea surface temperature (Virmani & Weisberg, 2006).

In South Florida, as elsewhere in cyclone-prone regions, several factors influence the persistence and productivity of mangrove communities where they are protected from coastal development. Davis et al. (2005) produced a conceptual model to explain mangrove forest productivity that emphasizes hydrological and phosphorous availability. Field data also show that reduced sheet flow from freshwater sources during the dry season (November – April) as well as limited tidal flushing in mangrove and other halophytic communities may lead to collapse of wetland vegetation and wetland substrates, including mangroves (Wanless & Vlaswinkel, 2005). A number of studies have also examined the influence of tropical storms and hurricanes on mangrove community resilience and resistance to these major disturbance events (Baldwin et al., 2001; Davis et al., 2004; Krauss, et al. 2005; Piou et al., 2006). Owing to their position along coastal margins, hurricane-related mortality in mangrove forests may be high (> 30 percent); however, mangrove trees seem to possess significant resilience if storm frequencies are sufficiently low to allow regeneration through re-sprouting and seedling growth (Baldwin et al., 2001; Piou et al., 2006).

Application of optical remote sensing to study mangrove environments has increased over the past decade (Blasco et al., 1998; Diaz & Blackburn, 2003; Murray et al., 2003; Vijay et al., 2005; Thampanya et al., 2006). Remote sensing studies have revealed much about mangrove ecosystems including canopy heights, leaf densities or leaf area index (LAI), deforestation, erosion, and pollution and other anthropogenic changes (Blasco et al., 1998). A number of remote sensing studies have employed multispectral satellite data from SPOT and Landsat, which provide resolutions appropriate for studying changes in mangrove canopy extent and condition. More recently, higher spatial and spectral resolution systems have been used for analysis of within-stand variability and changes (Held et al., 2003; Hirano et al., 2003; Vaiphasa et al., 2005). However, the cost of using new high-resolution systems remains high and these systems tend to provide limited spatial and temporal coverage relative to medium-to-low spatial resolution systems.

Empirical work by Green et al. (1997; 1998) and Kovacs et al. (2004; 2005) produced linear relationships between mangrove leaf area index and the normalized difference vegetation index or NDVI. These studies revealed that LAI of mangroves varies directly with NDVI and typically ranges from 2.5 – 4.0 in South Florida and the Caribbean. Because LAI is an important canopy parameter that relates to carbon uptake, net primary productivity, and surface energy balance (Myneni et al., 1997), much recent research has emphasized its global estimation using non-destructive, remotely sensed methods (Morisette et al., 2006; Soudani et al., 2006; Steinberg et al., 2006). Despite some success in estimating LAI for some ecosystems, global LAI estimation remains challenging for wetland environments where near infrared canopy reflectance is strongly attenuated by water present in the background. Furthermore, if many elements of the canopy and background optics are known or can be estimated, the use of appropriately parameterized radiative transfer models may help to overcome some of the site and time-specific limitations imposed by empirical formulae derived from intensive field studies (Myneni et al., 1997).

Study Area
The subtropical southwestern coast of Florida has a very gentle westward slope into the Gulf of Mexico and is in a leeward, low-energy setting with respect to prevailing easterly Trade Winds. According to Wanless & Vlaswinkel (2005), in times of sea level stability, colonization by mangrove wetlands has extended the coast significantly seaward. Currently, the region is characterized by low coastal ridges of marl and sand; vast mangrove, transitional, and freshwater wetlands; saline to freshwater lakes and lagoons; and deeply penetrating tidal channels extending from the freshwater Everglades drainage to the sea. Rapid changes are presently occurring in the coasts, coastal bays, and saline to freshwater wetlands of south Florida in response to sea-level rise over the past 70 years (Wanless & Vlaswinkel, 2005). A 24 cm relative rise in sea level in southwest Florida over the past 75 years (a rate 8 times that over the past 2,400 years) has partly inundated the low-lying marl ridges and caused coastal erosion, inundation, and loss of coastal barriers, and drastically modified coastal wetlands. This rise has destabilized the coastal mangrove communities and resulted in significant transgression, including storm erosion of mangrove coastlines (to 300 m), storm-initiated loss within mangrove forests, and landward expansion (to 1 km) of the red, black and white mangrove ecotones. Furthermore, interior freshwater wetland communities and soils are collapsing and evolving into broad, shallow, open-water areas as they become salt-stressed by rising sea level (Wanless & Vlaswinkel, 2005). Figure 1 shows the location of the study area with the 2005 hurricane tracks superimposed.

Figure 1: Location of the study area in southwest Florida showing an overlay of SPOT false color imagery and the two hurricane tracks (Hurricanes Katrina and Wilma) over the study region in 2005.

Data and Methods
The GeoSAIL model (Huemmrich, 2001) was used to estimate LAI from red reflectance and this estimate (equation in Figure 3) was applied to SPOT band 2 imagery to produce maps of mangrove LAI before and after the 2005 storms1. SPOT 20m imagery was acquired from the Center for Southeast Tropical Advanced Remote Sensing (CSTARS) at the University of Miami (see . Figure 2 shows SPOT imagery collected before the hurricane season in March 2005 and three days after landfall of Hurricane Wilma on 27 October 2005.

Figure 2. SPOT 20m false color imagery collected prior (2 March) to the 2005 hurricane season (left) and just after Hurricane Wilma (27 October), the last hurricane to affect south Florida that season (right).

GeoSAIL combines the scattering from arbitrarily inclined leaves (SAIL) model with a geometric model to simulate canopy spectral reflectance for discontinuous canopies. Mangroves were parameterized in the model as cylinders distributed over a plane. Spectral reflectance and transmittance of mangrove trees were calculated from the SAIL model to determine the reflectance of the three components of the geometric model: illuminated canopy, illuminated background, and shadowed background. SPOT band 2 images were atmospherically corrected using the Atcorr2 model supplied in the ERDAS IMAGINE software. GeoSAIL simulations were run using different assumptions and literature-based values applicable to mangrove canopies. Background reflectance was estimated by sampling atmospherically corrected band 2 values for mudflats and other mangrove substrates exposed in the SPOT 20m imagery. In addition, linear relationships presented by Green et al. (1997) and Kovacs et al. (2004) were used to generate maps of LAI based on SPOT NDVI imagery. These three different 20m products were subsequently compared to 1km MODIS LAI (MOD 15A2) imagery obtained from https://modis.gsfc.nasa.gov/data/dataprod/index.php.


Figure 3 shows GeoSAIL simulation results in which background reflectance was set at 0.06. The equation shows the fit using a logarithmic function, which was used as the basis for generating pre- and post-hurricane maps of LAI from SPOT band 2 imagery (right panel in Fig. 4).

Figure 3. GeoSAIL simulations used to estimate pre- and post-hurricane LAI of mangrove communities affected by 2005 hurricanes in southwest Florida.

Figure 4 shows the MODIS LAI for November 2005 and the post-hurricane SPOT-based LAI estimation derived using the fitted relationship displayed in Fig. 3. Figure 4 reveals general qualitative agreement between the MODIS LAI and GeoSAIL LAI.

Figure 4. Post-hurricane LAI: 1-km MODIS LAI (left) for November 2005 compared with SPOT LAI (right) from 27 October, 2005.

Table 1. Study area mean and standard deviation (in parentheses) of different LAI estimates derived from empirically established observations (Green et al., 1997; Kovacs, et al. 2004), the GeoSAIL approach shown in Figure 2, and MODIS (MOD15A2 product).

Discussion and Conclusions
All four LAI products (MODIS, GeoSAIL, Kovacs et al., 2004, and Green et al., 1997) contained anomalous values. MODIS and GeoSAIL approaches resulted in high values in some coastal pixels, in some cases LAI > 25. This may have been caused by the inclusion of open water (i.e., mixed pixels) and mangrove canopy. The results based on Kovacs et al. (2004) also produced negative values in the March and October images, with quite extensive negative values in the latter. This suggests that the Kovacs et al. (1997) equation does not apply well outside the degraded mangrove habitat in Mexico where nearly one-third of the trees were either dead or in a degraded state. Overall, the empirical formula from Green et al. (1997) produced the fewest negative pixels in the October 2005 LAI image. It should be noted that all anomalous values (i.e., LAI 8) were excluded in the calculations of the means and standard deviations shown in Table 1. Further, the MODIS LAI estimates pre- and post-hurricane appeared most consistent with published literature values for mangrove LAI in our region.

Table 1 shows mean LAI estimates pre-hurricane season, which ranged from 1.308 from the Kovacs et al. (2004) equation to 5.480 for the Green et al. (1997) equation. Relative to literature values, the GeoSAIL approach appeared to underestimate pre-hurricane mangrove LAI; whereas, MODIS LAI may have overestimated pre-hurricane LAI. The range of estimates narrowed for post-hurricane LAI, although the Kovacs et al. (2004) equation produced anomalously low LAI estimates. Field observations made four months after the passage of Hurricane Wilma revealed that the hurricanes produced partial-to-complete defoliation and much damage to woody canopy components. Some of observed damage may have been due to the storm surge from Hurricane Wilma that exceeded two meters along parts of the coastal zone (Pasch et al., 2006). This level of observed damage is consistent with the LAI changes shown in Table 1.

Linear fits applied to scatter plots (not shown) indicated modest agreement (r = 0.53) between one empirical formula (Green et al., 1997) and pre-hurricane MODIS LAI; while the GeoSAIL approach produced the best agreement between post-hurricane MODIS LAI and SPOT LAI (r = 0.69) among the different methods employed. The results suggest that in the absence of detailed field data, the GeoSAIL model combined with 20m multispectral satellite imagery provide an effective way to monitor LAI changes associated with the damage and subsequent recovery of mangrove canopies following tropical cyclones. Moreover, while further refinement of the GeoSAIL-based approach is needed, LAI estimates derived from 20m SPOT imagery possess clear advantages over 1km MODIS products, which mask local variation and spatial detail that might otherwise be revealed in SPOT 20m multispectral imagery.


  • Baldwin, A., Egnotovich, M., Ford, M. & Platt, W. (2001). Regeneration in fringe mangrove forests damaged by Hurricane Andrew. Plant Ecology, 157, 149-162.
  • Blasco, F.T., Gauquelin, T., Rasolofoharinoro, M., Denis, J., Aizpuru, M., & Caldairou, V. (1998). Recent advances in mangrove studies using remote sensing data. Marine and Freshwater Research, 49, 287-296.
  • Davis S.E., Cable, J.E., Childers, D.L., Coronado-Molina, C., Day, J.W., Hittle, C.D., Madden, C.J., Reyes, E., Rudnick, D. & Sklar, F. (2004). Importance of storm events in controlling ecosystem structure and function in a Florida gulf coast estuary. Journal of Coastal Research, 20, 1198-1208.
  • Davis, S.M., Childers, D.L., Lorenz, J.J., Wanless, H.R., & Hopkins, T.E. (2005). A conceptual model of ecological interactions in the mangrove estuaries of the Florida Everglades. Wetlands, 25, 832-842.
  • Diaz, B.M. & Blackburn, G.A. (2003). Remote sensing of mangrove biophysical properties: evidence from a laboratory simulation of the possible effects of background variation on spectral vegetation indices. International Journal of Remote Sensing, 24, 53-73.
  • Green, E.P., Mumby, P.J., Edwards, A.J., Clark, C.D., & Ellis, A.C. (1997). Estimating leaf area index of mangroves from satellite data. Aquatic Botany, 58, 11-19.
  • Green, E.P., Clark, C.D., Mumby, P.J., Edwards, A.J., & Ellis, A.C. (1998). Remote sensing techniques for mangrove mapping. International Journal of Remote Sensing, 19, 935-956.
  • Held, A., Ticehurst, C., Lymburner, L., & Williams, M. (2003). High resolution mapping of tropical mangrove ecosystems using hyperspectral and radar remote sensing. International Journal of Remote Sensing, 24, 2739-2759.
  • Hirano, A., Madden, M., & Welch, R. (2003). Hyperspectral image data for mapping wetland vegetation. Wetlands, 23, 436-448.
  • Hogarth, P.J., (1999). The Biology of Mangroves. Oxford and New York: Oxford University Press, 228 pp.
  • Kovacs, J.M., Flores-Verdugo, F., Wang, J.F., & Aspden, L.P. (2004). Estimating leaf area index of a degraded mangrove forest using high spatial resolution satellite data. Aquatic Botany, 80, 13-22.
  • Kovacs, J.M., Wang, J.F., & Flores-Verdugo, F. (2005). Mapping mangrove leaf area index at the species level using IKONOS and LAI-2000 sensors for the Agua Brava Lagoon, Mexican Pacific. Estuarine Coastal and Shelf Science, 62, 377-384.
  • Krauss, K.W., Doyle, T.W., Twilley, R.R., Smith III, T.J., Whelan, K.R.T., & Sullivan, J.T. (2005). Woody debris in the mangrove forests of South Florida. Biotropica, 37, 9-15.
  • Morisette, J.T., F. Baret, et al. (2006). Validation of global moderate-resolution LAI products: A framework proposed within the CEOS Land Product Validation subgroup. IEEE Transactions on Geoscience and Remote Sensing, 44, 1804-1817.
  • Muller, R.A. & Stone, G.W. (2001). A climatology of tropical storm and hurricane strikes to enhance vulnerability prediction for the Southeast US coast. Journal of Coastal Research, 17, 949-956.
  • Murray, M. R., Zisman, S.A., Furley, P.A., Munro, D.M., Gibson, J., Ratter, J., Bridgewater, S., Minty, C.D., & Place, C.J. (2003). The mangroves of Belize Part 1. distribution, composition and classification. Forest Ecology and Management, 174, 265-279.
  • Myneni, R.B., Nemani, R.R. & Running, S.W. (1997). Estimation of global leaf area index and absorbed PAR using radiative transfer models. IEEE Transactions on Geoscience and Remote Sensing, 35, 1380-1393.
  • Odum, W.E. & McIvor, C.C. (1990). Mangroves. In R.L. Myers, J.J. Ewel (Eds.), Ecosystems of Florida (pp. 517-548). Orlando: University of Central Florida Press.
  • Pasch, R.J., Blake, E.S., Cobb III, H.D., & Roberts, D.P. (2006). Tropical Cyclone Report: Hurricane Wilma, National Hurricane Center, Miami, FL, USA (available on line at https://www.nhc.noaa.gov/2005atlan.shtml).
  • Piou, C., Feller, I.C., Berger, U., & Chi, F. (2006). Zonation patterns of Belizean offshore mangrove forests 41 years after a catastrophic hurricane. Biotropica, 38, 365-374.
  • Saenger, P. & Snedaker, S.C. (1993). Pantropical trends in mangrove above-ground biomass and annual litterfall. Oecologia, 96, 293-299.
  • Soudani, K., C. Francois, le Maire, G., Le Dantec, V., & Dufrene, E. (2006). Comparative analysis of IKONOS, SPOT, and ETM+ data for leaf area index estimation in temperate coniferous and deciduous forest stands. Remote Sensing of Environment, 102, 161-175.
  • Steinberg, D.C., Goetz, S.J. & Hyer, E.J. (2006). Validation of MODIS F-PAR products in boreal forests of Alaska. IEEE Transactions on Geoscience and Remote Sensing, 44, 1818-1828.
  • Thampanya, U., Vermaat, J.E., Sinsakul, S., & Panapitukkul, N. (2006). Coastal erosion and mangrove progradation of Southern Thailand. Estuarine Coastal and Shelf Science, 68, 75-85.
  • Vaiphasa, C., Ongsomwang, S., Vaiphasa, T., & Skidmore, A.K. (2005). Tropical mangrove species discrimination using hyperspectral data: A laboratory study. Estuarine Coastal and Shelf Science, 65, 371-379.
  • Vijay, V., Biradar, R.S., Inamdar, A.B., Deshmukhe, G., Baji, S., & Pikle, M. (2005). Mangrove mapping and change detection around Mumbai (Bombay) using remotely sensed data. Indian Journal of Marine Sciences, 34, 310-315.
  • Virmani, J.I. & Weisberg, R.H. (2006). The 2005 hurricane season: An echo of the past or a harbinger of the future? Geophysical Research Letters, 33, L05707, doi:10.1029/2005GL025517.
  • Wanless, H.R. & Vlaswinkel, B. (2005). Coastal landscape and channel evolution affecting critical habitats at Cape Sable, Everglades National Park, Florida. Final Report to Everglades National Park, United State Department of the Interior, Homestead, FL, USA.