When strong forcing fails to raise turbidity: a regime-based diagnosis of state dependence in the Minjiang River Estuary from Sentinel−2 (2017–2025)

Understanding how turbidity in river-delta estuaries responds to short-term atmospheric variability is central to navigation, water quality, and ecosystem risk assessments. A widespread working assumption in many impact and hazard frameworks is that forcing–response mappings are approximately stationary: similar atmospheric forcing should, on average, yield comparable turbidity responses. Here we test this assumption for the Minjiang River Estuary (SE China) using a monthly Sentinel−2 record spanning January 2017–November 2025, quantifying turbidity variability with two complementary satellite proxies: (i) the areal fraction of persistently high turbidity (Afrac) and (ii) an intensity metric based on the upper tail of turbidity within the region of interest (Ip90). Breakpoint analysis identifies three regimes: R1 (Jan 2017–Sep 2022; n=67 months), an intermediate baseline state; R2 (Oct 2022–Feb 2024; n=17), an amplified and more variable state; and R3 (Mar 2024–Nov 2025; n=21), a suppressed state with markedly reduced variance. Critically, coastal pressure anomalies during R2 and R3 are statistically indistinguishable (Mann–Whitney p=0.7027; Cliff’s δ=0.076), yet turbidity anomalies collapse from elevated values in R2 (frac′ 0.301 ± 0.247) to near−zero in R3 (−0.011 ± 0.079), implying a regime-dependent change in response gain rather than weakened forcing. Phase−space diagnostics further reveal hysteresis, consistent with history dependence and sediment “memory” in the resuspendable bed supply. Together, these results provide satellite evidence that estuarine turbidity sensitivity can reorganize abruptly, challenging stationary assumptions commonly used in delta risk assessment and motivating monitoring approaches that track both external forcing and internal state variables.