Date of Graduation

7-2026

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Environmental Dynamics (PhD)

Degree Level

Graduate

Department

Environmental Dynamics

Advisor/Mentor

Ungar, Peter

Committee Member

Peter, Brad

Second Committee Member

Vogel, Erin

Third Committee Member

Fernandes, Katie

Keywords

NDVI; Orangutan; PDSI; Peat Swamp; Random Forest; Wildfire

Abstract

Tropical peat swamp forests are globally significant carbon reservoirs and biodiversity strongholds, yet they face escalating threats from climate change, fire, and anthropogenic disturbances. This dissertation investigates the interwoven dynamics of climate, fire, forest resilience, and orangutan survival in Central Kalimantan, Indonesia, a region which was transformed by the failed Mega Rice Project and recurrent catastrophic wildfires for decades. Using a multi-scale approach that integrates remote sensing time-series analysis (2001–2024),nlandscape-scale disturbance mapping (2005–2025), and fine-scale movement ecology, this research tests three core hypotheses: (1) moisture availability is the dominant climatic driver of wildfire activity, operating through non-linear thresholds; (2) post-fire forest recovery is limited and strongly mediated by disturbance type and frequency; and (3) the predictive power of remotely sensed habitat metrics for orangutan step selection is determined by spatial resolution. Chapter 1 demonstrates that precipitation and the Palmer Drought Severity Index (PDSI) are the principal correlates of fire activity. Months with precipitation below 209 mm show a 19.9-fold increase in burned area, and drought conditions (PDSI ≤ -1) show an 8.9-fold increase, confirming threshold-driven behavior. Temperature acts only as a contextual amplifier under pre-existing moisture deficit. Long-term trends indicate significant wetting and declining burned area, suggesting that episodic ENSO-driven droughts, not secular climate change, drive severe fire years. Chapter 2 quantifies forest disturbance and recovery over 20 years. Anthropogenic disturbances is dominated by urbanization (41.7%), logging (40.2%), and fire (18.1%). Only 34.5% of disturbed area shows recovery, with logged areas recovering best (42.8%), fire-affected areas intermediately (31.9%), and urbanized areas minimally (12.6%). Repeated burning reduces recovery, indicating eroding resilience. Chapter 3 provides the first direct test of NDVI as apredictor of fine-scale orangutan movement. Using step-selection functions on 151,668 steps from 60 individuals, forest cover at 30-m resolution is a consistent positive predictor, while MODIS NDVI at 250-m resolution contributes no significant predictive power for males and only a weak effect for females. Seasonal moisture cycling, not spatial structural variation, dominates NDVI signals. Females show stronger forest cover selection than males, a sex difference detectable only at finer resolution. Together, these findings reveal a cascade from climatic thresholds to fire regimes to degraded forest structure, which in turn shapes orangutan movement ecology. Resolution is a substantive parameter in habitat selection analyses. Effective conservation requires threshold-based fire early warning, hydrological restoration to break fire-recovery feedback, and fine-resolution habitat metrics to guide primate conservation in degraded tropical forests.

Download

Share

COinS