Long-term recovery dynamics determined by the degree of the disturbance - Ten years tracking of aquatic macroinvertebrate recolonisation after an industrial disaster (Red Sludge Disaster, Hungary).

A ten-year-long examination of macroinvertebrate community recovery was conducted following a catastrophic spill of highly alkaline red sludge (pH >13) into lowland streams. Our primary objective was to compare recovery patterns after coarse- and fine-grain disturbances, focusing on two aspects: i) trend analysis to reveal long-term changes of six community parameters, and ii) variation analyses to assess parameter changes over time. We conducted statistical analysis on long-term data series of macroinvertebrates obtained from quantitative samples collected at four sections with varying degrees of disturbance along the impacted stream sections. We developed a comprehensive theoretical framework comprising a series of sequential phases: Ramp-up, Overshoot, and Oscillation Phases. i) A trend analysis revealed that disturbances show a gradual recovery pattern, while variance analyses showed an asymptotic convergence to an equilibrium. ii) Evaluating these trends across phases unveiled that the initial recovery phase exhibited a steep trajectory, lasting 4-9 months, irrespective of disturbance severity. Coarse-grain disturbances induced a remarkable Overshoot phenomenon across all community metrics. The more severe the disturbance, the greater the height and duration of the Overshoot. Our results suggest that the presence or absence of Overshoot can serve as an indicator for coarse-grain disturbances in the context of large and infrequent disturbances (LID). The entire recovery process lasts for 2.5-3 years irrespective of the severity of the LID. In conclusion, a minimum survey duration of two and half years is deemed imperative to capture the phases of recovery, and changes associated with LID are not expected to extend beyond the three-year threshold. The theoretical framework, including Overshoot parameters, may assist future studies in comparing recovery patterns of different LID types. Furthermore, our theoretical framework is likely to be applicable to other groups of organisms given a sufficiently long monitoring of recovery, influenced also by the length of reproductive cycles.

Leaf litter decomposition in Torna stream before and after a red mud disaster.

The aim of the study was to estimate the breakdown of the allochthonous litter in an artificial stream running in an agricultural area and compare it with the same values following a toxic mud spill into the same stream. Litter bags were filled with three types of leaves (Quercus robur, Populus tremula and Salix alba) and placed to the bottom of the river. Ergosterol was used to detect fungal biomass. We supposed the absence of fungi and the retardation of leaf litter decomposition. Only pH and conductivity increased significantly. Leaf mass loss after the catastrophe was much slower than in 2009 and the decay curves did not follow the exponential decay model. Prior to the catastrophe, leaf mass loss was fast in Torna, compared to other streams in the area. The reason is that the stream is modified, the bed is trapezoid and covered with concrete stones. Fungal biomass was lower, than in the pre-disaster experiment, because fungi did not have enough leaves to sporulate. Leaf mass loss followed the exponential decay curve before the disaster, but after that it was possible only after a non-change period.