B116-0021
Structural predictors of solar-induced chlorophyll fluorescence (SIF) at intact and degraded forests in the Brazilian Amazon

Wednesday, 16 December 2020
Poster
Ekena Rangel Pinagé, University of Technology Sydney, School of Life Sciences, Ultimo, NSW, Australia; Oregon State University, College of Forestry, Corvallis, United States, David M Bell, USDA Forest Service, Corvallis, OR, United States, Marcos Longo, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, United States, Sicong Gao, University of Canberra, Centre for Applied Water Science, Canberra, Australia; CSIRO Land and Water Canberra, Canberra, Australia, Carlos Alberto Silva, University of Florida, School of Forest Resources and Conservation, Gainesville, FL, United States, Michael Keller, USDA Forest Service, International Institute of Tropical Forestry, Rio Piedras, PR, United States; Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, United States, Jean Pierre Ometto, INPE National Institute for Space Research (INPE), Earth System Science Center, Sao Jose dos Campos, Brazil and Alfredo R Huete, University of Technology Sydney, Faculty of Science, Ultimo, NSW, Australia
Abstract:
Tropical forest degradation (e.g., selective logging and fires) alters forest structure and function, increases structural heterogeneity and may affect the forest’s ability to efficiently harvest light for photosynthesis. In this study, we aimed to quantify the contribution of structural variables to solar-induced fluorescence (SIF) measured over intact and degraded forests in the southern Brazilian Amazon. We paired TROPOMI SIF observations with forest structural attributes derived from small-footprint airborne lidar data over intact and degraded forest areas. Multiple regression models were then developed with stepwise selection including SIF predictors from four broad aspects of forest structure (canopy height distribution, foliar distribution, horizontal complexity and vertical complexity) from lidar metrics. The model fitted with canopy height predictors showed the best performance (adjusted R2 = 0.44), and included the mean and standard deviation of height returns, and standard deviation of 25th (sdP25) and 95th (sdP95) height percentiles. The best performance of a single variable was achieved by the standard deviation of the vertical foliage profile (sdVFP), which explained about 24% of SIF variability. The inclusion of sdP25 and sdP95 metrics in the best model highlights the importance of understory and overstory heterogeneity caused by fires and logging to SIF estimates. We also found that older burned areas, that are most likely actively regenerating, showed higher observed and predicted SIF values, while intact and logged forests showed mixed ranges of SIF values. The increasing light levels reaching the understory of degraded forests, the decreasing canopy crown shadowing, and compositional shifts towards fast-growth, high photosynthetic-capacity species may be causing these SIF increases. Our findings suggest that disturbance and subsequent recovery mediate productivity in tropical forests, and advance the investigation of the mechanisms driving productivity in degraded tropical forest landscapes.
<< Previous Abstract | Next Abstract >>