Print PagePrecipitation, temperature, and other factors affect plant and ecosystem processes at multiple time scales, but a common assumption is that conditions at a given time directly affect processes at that same time period. Recent work in pulse-driven, semiarid systems shows that antecedent water availability averaged over the past several days to weeks can be just as or more important than current water status, and precipitation and temperature patterns of past seasons or years can also impact ecosystem functioning. However, we lack an analytical framework for quantifying the importance of and time-scale over which past conditions affect current processes. We address this by exploring the ecological “memory” of plant and ecosystem productivity, where memory is used as a metaphor to describe the time-scales over which antecedent conditions affect productivity. Existing approaches for incorporating antecedent effects arbitrarily select the integration period (e.g., past 2 weeks) and the relative importance of past conditions (e.g., assign equal weights to past events). Conversely, we utilize a hierarchical Bayesian approach to integrate field data with simple process models to estimate parameters describing the duration of the memory (integration period) and the relative importance of past events (weights) to this memory.
Results/Conclusions
We apply our approach to understand the memory of net primary productivity (NPP) and tree growth (ring widths). NPP data were extracted from Lauenroth and Sala (1992) for the shortgrass steppe; tree-ring data were obtained from the Tree-Ring Data Bank for pinyon pine. Precipitation received during the current and past two years explained 82% of the variation in NPP, whereas current year precipitation alone only explained 45%. Likewise, current and antecedent precipitation and temperature explained 72% of the variation in ring widths, while current conditions alone only explained 50%. However, the precipitation and temperature tree-ring memories differed. While the current and past two years accounted for 85% of the precipitation memory, temperature conditions during the current year were not important; only temperature conditions during the previous two years were important. These dissimilar memories likely reflect differential impacts of water and temperature on growth processes. Conditions experienced >3 years ago had little influence on NPP and tree growth, and it remains to be tested if this integration period is broadly applicable to annual productivity processes in semiarid systems. This study highlights the importance of understanding the temporal scales over which environmental factors influence productivity, and our approach to quantifying ecological memory lends insight into potential mechanisms.
