Research Professor University of Tennessee Knoxville Knoxville, Tennessee, United States
Abstract: Succession, or the development of an ecosystem subsequent to disturbance, is a process found in every system. Here we examine previous research on primary and secondary succession, occuring on mountainous landslides and debris avalanches (volcanically triggered slides with greater than 50% rock content). These disturbances belong to a class of disturbance that could be classified as infrequent and severe. We wanted to identify what constrains succession in these landscapes, whether generalizable trajectories exist, and how long succession might be expected to take.
We conducted a systematic review of previous research. Using three databases and ten search terms, we collated all primary research articles on vegetative succession that has occurred subsequent to debris avalanches and mountainous landslides published before 9 Dec 2022. Titles and abstracts were reviewed to determine if they matched search criteria. Some 18 (Web of Science)+118 (Jstor)+195 (Google Scholar) = 331 potential titles were found in the initial sweep. Closer review found that ~71% of the articles met the search criteria.
We found similarities in patterns of species accumulation and cover development across mountains. Linear models over time were generally a good fit for accumulation of number of species per area with a mean R2 of .82 (SE .05). Debris avalanches tended to have a positive slope with a mean accumulation rate of 1.52 (SE 0.65) species per year. In contrast to debris avalanches, landslides were smaller and tended to be rapidly colonized, and then a few species took over, leading to a loss of species over time at a rate of -1.88 (SE 0.52) species per year. An exponential increase in cover over time was a good fit for most successional time series, mean R2 .62 (SE .10). Where the fit to an exponential distribution was poor, the successional time series had reached a stable maximum cover. Landslides [.61 (SE .16)] and debris avalanches [.79 (SE .12)] appear to fit this assumption equally well (p-value .41). Only two studies covered a long enough time scale to access treeline depression, both on debris avalanches. In both cases tree line depression still existed after 400+ years. It was difficult to quantitatively access patterns in community change between studies due to differing methodologies. Some systems seem to reach stable community dynamics in terms of functional type in as little as 2 years, while others are still shifting considerably after 20 plus years.