Estimation of Water Stage Over Wetlands of South Florida Using TRMM Precipitation Radar Observations

Document Type

Conference Proceeding


Everglades are a critical component of the regional hydrological cycle in South Florida. Anthropogenic activities in this region have deteriorated the wetland ecosystem and efforts have been made to restore and preserve it. Seasonal and interannual changes in water stage result in saltwater intrusion and inhibit ecosystem conservation measures. Hence, there is a need to monitor water stage in wetlands. Microwave remote sensing with its sensitivity to surface characteristics provides an opportunity to measure changes in water stage from space. Spaceborne remotely sensed images can provide a comprehensive spatio-temporal distribution of water stage over an area thereby eliminating the need to monitor water stage separately at each measurement site. This research relates water stage measurements (ws) to Tropical Rainfall Measuring Mission Precipitation Radar backscatter (σ°). σ° response to partially exposed vegetation is used as the basis of the model. Variations in the water depth change the amount of exposed vegetation canopy that is reflected in the σ° measurements. An empirical linear model is developed that expresses ws in terms of σ°. The impact of vegetation on the model is studied by examining model performance over various landcovers. The ws model is applied to stage data on sites operated by South Florida Water Management District. Eleven year data (1998 to 2008) is used for this research. The model is calibrated using 75% of the time period of data to estimate model parameters. The model is validated using the remaining 25% of the time period. The estimated water stage measurements from the model are compared with observed measurements over different landcovers. The model performance is assessed by comparing correlation coefficient (R), root mean square error (rmse), and mean absolute error (mae) between observed and modeled water stage measurements. The model works reasonably well in the regions with tree heights greater than 5 m such as deciduous forest (R=0.58, rmse=0.58 ft, mae=0.45 ft), mixed forest (R=0.61, rmse=0.78 ft, mae=0.57 ft), and woodlands (R=0.56, rmse=0.65 ft, mae=0.54 ft). Other low lying landcovers such as cropland (R=0.27, rmse=0.63 ft, mae=0.48 ft) and closed shrubland (R=0.33, rmse=0.56 ft, mae= 0.43 ft) do not show significant performance of the model. This is because the vegetation in croplands and closed shrubland are submerged under water for most part of the year. The incident microwave radiations get specularly reflected from the water surface resulting in lesser backscatter. On the other hand, in areas with tall vegetation, the incident radiations get backscattered from the vegetation above the water surface exhibiting distinct response as a result of the amount of submergence. The modeled values of water stage compare well with the observed water stage in areas with tall vegetation. Thus microwave remote sensing can provide a comprehensive spatio-temporal distribution of water stage. This research provides a new insight into measurement of water stage using spaceborne remote sensing techniques.


Civil and Environmental Engineering | Civil Engineering | Environmental Engineering | Environmental Health and Protection | Environmental Monitoring | Environmental Sciences | Water Resource Management


Conference held: Providence, Rhode Island, United States, May 16-20, 2010


Use Find in Your Library, contact the author, or interlibrary loan to garner a copy of the item. Publisher policy does not allow archiving the final published version. If a post-print (author's peer-reviewed manuscript) is allowed and available, or publisher policy changes, the item will be deposited.