Theses Doctoral

Impacts of drought and cyclonic storms on vegetation and streamflow in Puerto Rico

Hall, Jazlynn

Forests are crucial for mitigating the effects of climate change and providing ecosystem services. Global deforestation trends coupled with changes in climate and disturbance regimes threaten the ability of forests to sequester carbon, provide timber, and regulate other ecosystem processes. My dissertation draws from principles in ecology, hydrology, and geography to identify the effects of forest disturbance from extreme events (i.e., hurricanes and droughts) on carbon sequestration and streamflow in Puerto Rico.

In Chapter 1, I utilized hydroclimatic time series, geospatial analysis, and statistical techniques to evaluate the relationships between forest cover and rainfall variability on streamflow regimes. I determined that (i) watersheds with high forest cover had consistently greater streamflow than deforested ones and (ii) during dry antecedent rainfall conditions, highly forested watersheds had higher streamflow than deforested ones, suggesting that increased hillslope infiltration and storage afforded by well-developed forest soils outweigh higher vegetation water demand in forests compared to other land cover types.

In Chapter 2, I combined field data from permanent forest plots with geospatial attributes, satellite imagery, and lidar point clouds to estimate forest damage extent and a loss in carbon sequestration across Puerto Rican forests after Hurricane María (2017). I also developed a random forest model to identify the biophysical and climatic drivers of spatial variation in forest damage from the hurricane. I found (i) a strong relationship between field-based estimates of aboveground biomass (AGB) loss after Hurricane María and the remotely sensed estimate of damage, (ii) 23% (10.44 ±2.33 Tg) of pre-hurricane forest AGB was lost, and (iii) storm-related rainfall was a strong predictor of forest damage, in addition to canopy height, maximum wind speeds, and soil moisture and water storage variables.

Finally, in Chapter 3, I quantified changes to streamflow in Puerto Rico in the months following Hurricane María and tested whether estimates of storm-related forest damage account for these differences. I found that (i) “fast” (i.e. quickflow) and “slow” (i.e., baseflow) streamflow components reacted differently to Hurricane María, (ii) magnitude of change in quickflow was positively correlated to the magnitude of forest damage from the storm, while changes to baseflow were unrelated to forest cover or damage, and (iii) under the same level of relative damage, watersheds with low forest cover had greater quickflow increases than deforested ones, suggesting that interactions between forest cover and forest damage modulate the effects of disturbance on quickflow in highly forested watersheds and exacerbate effects of disturbance in less forested watersheds.

Overall, the results from my dissertation inform management practices by quantifying vulnerability to changing climate and disturbance regimes, and highlight the importance of forests for regulating water provision and other ecosystem services.

Geographic Areas


  • thumnail for Hall_columbia_0054D_17184.pdf Hall_columbia_0054D_17184.pdf application/pdf 4.84 MB Download File

More About This Work

Academic Units
Ecology, Evolution, and Environmental Biology
Thesis Advisors
Uriarte, Maria
Defries, Ruth
Ph.D., Columbia University
Published Here
April 27, 2022