Spatial replication can best advance our understanding of population responses to climate

Abstract

Understanding the responses of plant populations dynamics to climatic variability is frustrated by the need for long term datasets that capture demographic responses to a range of climates. Here, we advocate for new studies that prioritize spatial over temporal replication, but without inferring the effect of temporal climatic gradients from spatial climatic gradients, as usually done in the so called space for time substitutions. Rather, we advocate to estimate the effects of climate by sampling replicate populations in locations with similar climate. We first use data analysis on spatial locations in the conterminous USA to assess how far apart spatial replicates should be from each other to minimize temporal correlations in climate. We find that spatial locations separated by 250 Km have moderate (0.5) correlations in annual precipitation. Second, we use simulations to demonstrate that spatial replication can lead to substantial gains in the range of climates sampled during a given set of years so long as the climate correlations between the populations are at low to moderate levels. Third, we use simulations to quantify how many spatial replicates would be necessary to achieve the same statistical power of a single-population, long-term data set under different strengths and directions of spatial correlations in climate between spatial replicates. Our results indicate that spatial replication is an untapped opportunity to study the effects of climate on demography and to rapidly fill important knowledge gaps in the field of population ecology.

Antecedent effect models as an exploratory tool to link climate drivers to herbaceous perennial population dynamics data

“…The predictive success of null models is a clear refutation of our initial expectation. We expected dataset length would positively affect the predictive ability of climatic variables by increasing total sample size, and by increasing the range of climatic conditions observed during the study (Compagnoni et al 2024). For the fact that climate had small predictive power even in datasets exceeding 30 years we propose three possible explanations.…”

“…We consider this spatial resolution more than sufficient, because climate anomalies are strongly correlated spatially. For example, in North America, to observe correlations of 0.5 between the annual anomalies of two sites, one needs to travel at least 300Km, for precipitation anomalies, and 1,400 Km, for temperature anomalies (Compagnoni et al 2024).…”

“…Our case study modeled reproduction in Astragalus tyghensis, an endemic legume plant species with an extremely restricted range encompassing arid lands of Western North America (Thorpe and Kaye 2008). This dataset has only nine years of data but five sites, which must have substantially improved the chance to detect climatic signals (Compagnoni et al 2024). Using this example, we first show how antecedent effect models identify monthly climatic drivers of demographic responses, and we discuss how these inferences can inform biological hypotheses.…”

A unified framework to identify demographic buffering in natural populations

“…We Although we are aware that not all possible temporal variation in demographic processes may have been expressed within this period, we assumed three or more transitions are enough to provide sufficient variation for population comparison. At least three contiguous time periods -a common selection criteria in comparative studies of stochastic demography (Compagnoni et al 2023) -also allowed to test and showcase our framework. Fortunately, several long-lived species, characterized by low variation in their demographic processes, were studied for a long time (e.g., some primates in our dataset have been studied for over 20 years - Morris et al 2011).…”