What question did this study set out to answer?

The aim is to identify climatic and anthropogenic factors contributing to mangrove forest disturbances over several decades.

June 3, 2026Open Access

Multi-Decadal Assessment of Climatic, Productivity, and Anthropogenic Drivers of Mangrove Disturbance

Key Points

The aim is to identify climatic and anthropogenic factors contributing to mangrove forest disturbances over several decades.
Analyzed multi-decadal Landsat time series (1990–2023) for mangrove mapping
Used logistic regression models to associate disturbance occurrences with climate and accessibility factors
Examined various proxies like NDVI, productivity, and accessibility gradients
Disturbance area and patch density increased significantly after 2010, mostly affecting small, dispersed patches (<3 ha).
Productivity anomalies contributed 51.6% to disturbance likelihood, with climate and accessibility factors at 26.6% and 21.8%, respectively.
Higher disturbance odds were linked to positive productivity anomalies (OR ≈ 1.31) and certain climatic conditions, like higher shortwave radiation.

Abstract

Mangrove forests provide critical ecosystem services, yet global declines have accelerated under combined anthropogenic pressure and climate variability. Identifying which environmental and accessibility gradients are most strongly associated with canopy disturbance is important for monitoring and management. Here, we analysed a multi-decadal Landsat time series (1990–2023) to map mangrove disturbance in Sarawak, Malaysia, using Landsat-based temporal segmentation (LandTrendr) of the Normalized Burn Ratio (NBR) and the Normalized Difference Moisture Index (NDMI). We then fitted bin-stratified logistic regression models to quantify associations between disturbance occurrence and event-year anomalies in climate (precipitation, shortwave radiation, minimum temperature, vapour pressure deficit), canopy condition proxies (kernel NDVI, kNDVI), satellite/model-derived productivity (net primary productivity, NPP), and anthropogenic accessibility (distance to roads, open water, mangrove edge, and built-up intensity). Disturbance area and patch density increased abruptly after 2010, with most mapped events being small (<3 ha) and spatially dispersed. In a model-based dominance analysis (dominant linear-predictor contribution by group), the largest share of disturbed pixels was most strongly associated with productivity/greenness anomalies (≈51.6% of the filtered disturbed footprint), followed by climate anomalies (≈26.6%) and anthropogenic accessibility (≈21.8%). Negative kNDVI anomalies were associated with substantially lower odds of disturbance (odds ratio ≈ 0.097), whereas positive NPP anomalies were associated with higher odds (OR ≈ 1.31). Brighter and drier-than-average conditions (higher shortwave radiation and positive precipitation anomaly, as defined here) corresponded to higher disturbance odds, while cooler night-time conditions were associated with lower odds. Among accessibility covariates, proximity to roads showed a strong association with higher disturbance probability, whereas built-up intensity was not significant after accounting for distance gradients. Overall, the results provide an association-based framework for prioritising monitoring and management by jointly considering climate variability, canopy-condition signals, and access corridors.

Read Full Paperexternally

Ask AI

Helpful

Bookmark

View Full Paper