Climate change and anthropogenic pressures are increasingly driving abrupt regime shifts (RSs) in natural ecosystems, leading to biodiversity loss and disruptions in their functioning. In systems with alternative stable states maintained by feedback mechanisms, such shifts are difficult to reverse due to hysteresis. These systems may show early warning signals (EWSs) (e.g. critical slowing down (CSD)), indicating ecological resilience loss prior to an RS. CSD manifests as a reduced rate of recovery from disturbances and can be detected through increasing temporal autocorrelation at-lag-1 (AC1) in vegetation indicator time series. While studies increasingly use remote sensing data (e.g. the Normalized Difference Vegetation Index) to evaluate CSD in natural ecosystems, most lack in-field validation and are conducted at broad spatial scales, limiting real-world applicability. Here, we assess whether EWSs precede RSs in Zambian wet Miombo woodlands by combining field-based vegetation data with remote sensing time series. We investigate (i) whether RSs in the Miombo woodlands can be detected using field data on vegetation structure and remote sensing data, (ii) the relationship between fire frequency and RSs, and (iii) if increasing AC1 precedes an RS. Our findings demonstrate the occurrence of RSs in Miombo woodlands and provide empirical evidence supporting the use of CSD as a predictive tool. Spatial models showed that both a high fire frequency and an increasing AC1 preceding the shift are significantly (p < 0.05) linked to RS occurrence. This study highlights the importance of integrating field and remote sensing data to monitor ecosystem transitions. By identifying resilience loss and RS occurrence, our approach offers valuable insights for conservation and restoration planning, emphasizing the need for adaptive fire management strategies to mitigate degradation and promote ecosystem recovery.

Read original article

Read original article