Analysis of Evaporation Variables in Open Water and Penman Monteith Evaporation Variables Using Ridge Regression [Case Study Forest Fire Model]
Year: 2025
Type: article
Abstract: Forest and land fires constitute a calamity influenced by actual evapotranspiration (act) derived from ERA5-land data, evaporation from open water (pev) derived from ERA5-Land data, and Penman-Monteith evaporation (et0). This study aims to scrutinize the characteristics, spatial and temporal aspects of these three variables, and assess the efficacy of pev and et0 in modeling forest fires through Ridge Regressions. Forest fire-related data, acquired via the Empirical Orthogonal Function method based on Singular Value Decomposition, will be employed. The analysis characteristic, temporal, spatial, and ridge regression assessments, with hotspot data serving as an indicator for forest and land fires. The findings reveal that the pev variable requires the fewest computation indicators. The characteristics of the act variable closely resemble those of the et0 variable. All three variables exhibit similar spatial and temporal patterns in forest fire-related data. The performance of the pev in modeling forest fires is better than the reference evapotranspiration variable in most local region analysis. This result highlights the potential applications of pev in modeling the general characteristics of fire events, making it robust across different regions. However, for modeling extreme events, et0 proves to be superior to other analyzed variables, particularly for events not included in the training data. Thus, et0 has higher potential performance in analysing and modeling future impact of climate changes related to fires risk rise due to more frequent drought condition.
Language: English
Topic: Fire effects on ecosystems
Subfield: Global and Planetary Change
Field: Environmental Science
Domain: Physical Sciences
SDG: Climate action
Open Access status: diamond
OpenAlex ID: W4406359576