Remote sensing of leaf area index as an indicator of ecosystem resilience to El Niño-driven climate variability in the UK

The El Niño-Southern Oscillation (ENSO) influences vegetation far beyond the tropical Pacific, but its imprint on temperate maritime vegetation, where the teleconnection acts indirectly through winter circulation, temperature and precipitation, has rarely been quantified at the site scale. We analysed monthly MODIS leaf area index (LAI; MOD15A2H, Collection 6.1) for 2001–2023 at four selected United Kingdom sites, a montane heath (Cairngorms National Park), a lowland pine plantation (Thetford Forest), a lowland heath-woodland mosaic (New Forest) and an upland blanket bog (Dark Peak, Peak District), together with the NOAA Oceanic Niño Index (ONI). We quantified lagged (0–12 months) and seasonal correlations between detrended LAI anomalies and the ONI with autocorrelation-corrected significance and 95% confidence intervals (CIs), and tracked a critical-slowing-down resilience indicator, the rolling lag-1 autocorrelation (AC1) of LAI anomalies, around the strong El Niño events of 2009–10 and 2015–16. LAI-ONI correlations were negligible at the montane, plantation and lowland mosaic sites (|r| ≤ 0.07) and weakly positive at the blanket bog (r = 0.17–0.21), and none remained significant after correction for multiple testing (all false-discovery-rate-adjusted p ≥ 0.36). Growing-season anomalies showed no significant relationship with the preceding winter ONI (r = −0.13 to 0.27; all p ≥ 0.2), but with 21–23 years the 95% CIs extend to r ≈ 0.6, so moderate associations cannot be excluded. ENSO was not detectably expressed in the winter North Atlantic Oscillation or in local winter-spring temperature and precipitation (all p ≥ 0.19). Post-event AC1 changes were not unusual, and apparent AC1 trends were not robust to serial dependence or data gaps; cloud and snow left 60% of months unobserved at the montane site, where AC1-based inference was unreliable. The dominant signal was a greening of the blanket bog (summer LAI +0.45 m 2 m −2 per decade; Kendall’s τ = 0.55, p < 0.001), consistent with peatland restoration. For this record and these four sites, no statistically significant ENSO imprint on vegetation structure was detected, suggesting that ENSO forecasts offer little site-scale early warning for such ecosystems, where local drivers should take priority; weak or moderate effects, and responses elsewhere in the United Kingdom, cannot be ruled out.