Search for: All records

Species' distributions and abundances are shifting in response to climate change. Most species harbor microbial symbionts that have the potential to influence these responses. Mutualistic microbial symbionts may provide resilience to environmental change by protecting their hosts from increasing stress. However, environmental change that disrupts these interactions may lead to declines in hosts or symbionts. Microbes preserved within herbarium specimens offer a unique opportunity to quantify changes in microbial symbiosis across broad temporal and spatial scales. We asked how the prevalence of seed-transmitted fungal symbionts of grasses (Epichloe endophytes), which can protect hosts from abiotic stress, have changed over time in response to climate change, and how these changes vary across host species' ranges. Specifically, we analyzed 2,346 herbarium specimens of three grass host species collected over the last two centuries (1824 -- 2019) for the presence or absence of endophyte symbiosis, and evaluated spatial and temporal trends in endophyte prevalence. We found that endophytes increased in prevalence over the last two centuries from ca. 25% prevalence to ca. 75% prevalence, on average, across three host species. We also found that changes in prevalence were associated with observed changes in seasonal climate drivers; notably increasing precipitation corresponding to each host species' peak growing season and changes in off-peak season variability in precipitation. Our analysis performed favorably in an out-of-sample predictive test with contemporary data, however we identified greater local-scale variability in endophyte prevalence in contemporary data compared to historic data, suggesting that model fusion may be an important step moving forward. Our results provide novel evidence for a cryptic biological response to climate change that may contribute to the resilience of host-microbe symbiosis through context-dependent benefits that confer a fitness advantage to symbiotic hosts under environmental change. 

We designed novel field experimental infrastructure to resolve the relative importance of changes in the climate mean and variance in regulating the structure and function of dryland populations, communities, and ecosystem processes. The Mean x Variance Experiment (MVE) adds three novel elements to prior designs (Gherardi & Sala 2013) that have manipulated interannual variance in climate in the field by (i) determining interactive effects of mean and variance with a factorial design that crosses a drier mean with increased (more) variance, (ii) studying multiple dryland ecosystem types to compare their susceptibility to transition under interactive climate drivers, and (iii) adding stochasticity to our treatments to permit the antecedent effects that occur under natural climate variability. This new infrastructure enables direct experimental tests of the hypothesis that interactions between the mean and variance of precipitation will have larger ecological impacts than either the mean or variance in precipitation alone. A subset of plots have soil moisture and temperature sensors to evaluate treatment effectiveness by addressing, How do MVE manipulations alter the mean and variance in soil moisture and temperature? And, how does micro-environmental variation among plots influence how much MVE treatments alter soil moisture profiles over three soil depths? This data package includes soil moisture and temperature sensor data from the Mean x Variance Climate experiment in the Desert grassland ecosystem at the Sevilleta National Wildlife Refuge, Socorro, NM. 

We designed novel field experimental infrastructure to resolve the relative importance of changes in the climate mean and variance in regulating the structure and function of dryland populations, communities, and ecosystem processes. The Mean - Variance Experiment (MVE) adds three novel elements to prior designs that have manipulated interannual variance in climate in the field (Gherardi & Sala, 2013) by (i) determining interactive effects of mean and variance with a factorial design that crosses reduced mean with increased variance, (ii) studying multiple dryland biomes to compare their susceptibility to transition under interactive climate drivers, and (iii) adding stochasticity to our treatments to permit the antecedent effects that occur under natural climate variability. This new infrastructure enables direct experimental tests of the hypothesis that interactions between the mean and variance of precipitation will have larger ecological impacts than either the mean or variance in precipitation alone. A subset of plots have soil moisture and temperature sensors to evaluate treatment effectiveness by addressing, How do MVE manipulations alter the mean and variance in soil moisture and temperature? And How does micro-environmental variation among plots influence how treatments alter soil moisture profiles over three soil depths? This data package includes sensor data from the Mean x Variance experiment in the Plains grassland ecosystem at the Sevilleta National Wildlife Refuge, Socorro, NM, which is dominated by the grass species Bouteloua gracilis (blue grama). 

This project was designed to understand the demographic effects of vertically transmitted fungal endophytes (Epichloë spp.) on their grass hosts. The experiment includes seven host-symbiont taxonomic pairs: Agrostis perennans - E. amarillans, Elymus villosus - E. elymi, Elymus virginicus - E. elymi or EviTG-1, Festuca subverticillata - E. starrii, Poa alsodes - E. alsodes, Poa sylvestris - E. PsyTG-1, Schedonorus arundinaceus - E. coenophiala. Experimental plots were established at the Indiana University Lilly-Dickey Woods Research and Teaching Preserve in south-central Indiana, USA in 2007. For each species, 5-10 plots were planted with naturally symbiotic (S+) hosts, and 5-10 plots were plated with hosts that were disinfected of fungal endophytes by heat treatment (S-). Over 15 years (2007-2022) we collected demographic data on the survival, growth, reproduction, and recruitment of all plants in all plots. Beginning in 2018 we also collected data on the locations of all plants in every plot. 

This study was initiated in 2004 by TEX Miller to quantify individual-level demographic rates of tree cholla cacti (Cylindriopuntia imbricata) at the Sevilleta National Wildlife Refuge, Socorro Co., NM. This data package spans 2004-2018, though the study is maintained through the present. In each year, censuses were conducted over several days in the last week of May and/or the first week of June. Individuals were marked with aluminum tags and tracked longitudinally through time, with annual measures of survival, size, reproductive output, and interactions with insect herbivores and mutualists (extra-floral nectar-feeding ants). For the years 2004-2008, plants were grouped into spatial blocks (T1, T2, T3) lacking defined boundaries. Beginning in 2009, the study design changed to 30m x 30m plots (Plots 1-8) with an entirely new cohort of individuals that does not overlap with the 2004-2008 data. From 2009 onward, plots were searched each year and new plants were added to the census as they were discovered. Seedlings typically require several years of growth before they are detectable in our census, so counts of new plants in a given year do not necessarily reflect recruitment in that year. The data frame is set up such that one row is a transition year (May of year t to May of year t+1) for one individual; this means that all but the first and last years’ observations occur twice, once as the start of a transition year and once as the end.