Spatial temporal analysis using Hierarchical Bayesian approach: effects of climate variability on primary production of Deciduous Forests of the Northeastern US
MetadataShow full item record
Extreme weather events including drought, heavy rain, and heat waves are occurring with higher frequency and can cause substantial impacts to human and natural systems, such as forestry, water resources, and human health. One potential impact is on the productivity of deciduous forest ecosystems, which cover 22% of the planet and are an important component of the global carbon cycle. Yet, the influence of extreme weather on temperate forest production is poorly understood. This research estimates the effects of extreme weather on Gross Primary Production (GPP, a measure of how much carbon the forest removes from the atmosphere) of temperate deciduous forests in the Northeastern United States using data from the FLUXNET monitoring network. I used Hierarchical Bayesian models to account for climate variability (anomalies) and short-term weather events (such as drought, heat) to quantify the effects of climatic variability on forest ecosystem productivity. I also explored different lagged effects of climate variability on GPP and found that GPP is sensitive to climatic variability at different temporal scales and that the effects vary seasonally. For example, heavy precipitation has a positive effect on NEP early in the growing season but switches to a negative effect in late summer. In addition, I found that incorporating a lagged effect into the model reveals improves model performance suggesting that some of the short term variability can affect GPP for several weeks. In summary, this research improves our understanding of how lagged effect influence the GPP prediction and how climatic variability (anomalies and events) drives changes in ecosystem productivity and, ultimately, the global carbon cycle.