Spatial variations in salinity stress across a coastal landscape using chlorophyll fluorescence and vegetation indices derived from hyperspectral imagery

Background/Question/Methods

Environmental stresses may temporarily affect the photosynthetic apparatus, especially photosystem II (PSII) before irreversible morphological damage is apparent. Non-invasive remote sensing techniques, such as chlorophyll fluorescence and plant reflectance, have been developed to monitor plant stress and photosynthetic status, and to detect and predict changes in the natural environment.  Stress detection prior to visible damage could permit quick and accurate assessment of the physiological response to environmental stress at the landscape level, revealing spatial variation in the physiological status of plants and the environment.  Our objective was to investigate spatial variations in stress in the dominant woody vegetation of a Virginia barrier island using chlorophyll fluorescence and landscape-level reflectance imagery.  Field measurements of fluorescence, tissue chlorides and relative water content (RWC) were made on Myrica cerifera and Iva frutescens at the same time as airborne hyperspectral images were collected during a severe drought (2007) and compared to an extremely wet summer (2004) .  
Results/Conclusions

RWC and the water band index (WBI₉₇₀) indicated that water stress did not vary across the island.  In contrast, there were significant differences in tissue chlorides across sites for both species.  Using the physiological reflectance index (PRI), we were able to detect salinity stress across the landscape.  For M. cerifera PRI did not differ between wet and dry years, while there were differences in PRI during the two years for I. frutescens, possibly related to flooding during the wet year.  There was a positive relationship between PRI and DF/F′_m for M. cerifera (r² = 0.79) and I. frutescens (r² = 0.72).  The normalized difference vegetation index (NDVI), the chlorophyll index (CI) and WBI₉₇₀ were higher during the wet summer for M. cerifera, but varied little across the island.  CI and WBI₉₇₀ were higher during 2004 for I. frutescens, while there were no differences in NDVI during the two years.  PRI was not significantly related to NDVI, suggesting that the indices are spatially independent.  These results suggest that PRI may be used for early identification of salt stress that may lead to changes in plant distributions at the landscape level as a result of rising sea-level.  Comparsions between the two species indicate that variations in PRI and other indices may be species specific.