Recession constants are not constant: the impacts of multi-annual drought on recession behaviour and catchment storage.

We studied changes in the recession behaviour of catchments that experienced multi-annual drought conditions to explore their relationship with previously observed drought-induced shifts in catchments’ hydrological response (as measured by annual rainfall-runoff relationships). We found that recession behaviour can change significantly during persistent drought, highlighting the role of subsurface storage dynamics and catchment conductivity in determining catchments’ hydrological response to prolonged dry periods.

Analysis of streamflow recessions is commonly used to characterise catchment behaviour, and to explore the role that catchment storage plays in streamflow production. Recession techniques characterise average catchment behaviour over sufficiently long periods of recorded data, making them generally unsuitable for analysis of nonstationary hydrological conditions. Nevertheless, in the context of long-term drought, where nonstationarity arises over decadal periods, analysis of changes in catchment recession behaviour over time is possible. For this study, we consider nonstationarities in the catchment-level annual rainfall-runoff relationships induced by prolonged drought. These have been observed during multi-annual droughts worldwide and often persist long after the end of the dry spell. In this context, recession analysis is a useful tool to study the effects of persistent drought on catchment processes.

We applied recession analysis methods to assess changes in average recession behaviour of catchments affected by multi-annual drought in Sout-Eastern Australia. We compare recession behaviour before the drought to a 10-year period straddling the end of the drought. We focussed on how significant changes in recession behaviour over time correlate to drought-induced shifts in annual rainfall-runoff relationships. We apply two distinct methods, drawn from the vast methodological literature on recession analysis. These were chosen specifically for their different data requirements (hourly or daily) and approaches to recession analysis (one based on a master recession curve and the other based on recession plots).

Despite the differences, results from both methods are consistent. We found that recession behaviour changed significantly in the majority of the catchments studied, with recessions becoming faster late in the drought compared to the pre-drought period. These changes, in particular, affected catchments that were shown to exhibit significant shifts in rainfall-runoff relationship during the extended drought. Conversely, in the catchments whose rainfall-runoff relationship had remained stable, the changes in recession behaviour are much smaller and largely limited to the low-flow portion of the recession curve. This suggests that the widespread increase in recession rates observed in shifted catchments is only in small proportion attributable to increased evaporative demand (which is comparable between the two sets of catchments) and is instead likely caused by a combination of decreased connectivity between catchment surface water and subsurface storage and increased transmission losses through the streambed.