Bayesian spatial modeling of moose count data: increasing estimator efficiency and exploring ecological hypotheses

Moose management throughout much of Alaska and Canada relies on aerial count data, which are commonly collected using the geospatial population estimator (GSPE) protocol. The GSPE uses a model-based analytical approach and finite-population block kriging to estimate abundance from a collection of sampled survey units. Widespread implementation and well-documented analytical software have resulted in reliable estimates of moose abundance, density, and composition across a large proportion of their range. Analysis is conducted almost exclusively using the GSPE software, which fits a fixed model structure to data collected within a single year. The downside of this approach to analysis is that the fixed model structure is inefficient for estimation, leading to more field effort than would otherwise be necessary to achieve a desired level of estimator precision. We developed a more easily modified and flexible Bayesian spatial general additive model approach (BSG) that accommodates spatial and temporal covariates (e.g., habitat characteristics, trend), multiple survey events, prior information, and incomplete detection. Using a series of 6 GSPE surveys conducted in Yukon-Charley Rivers National Preserve, Alaska, USA, from 2003–2019, we established the equivalence of the 2 approaches under similar model structures. We then extended the BSG to demonstrate how a more comprehensive approach to analysis can affect estimator precision and be used to assess ecological relationships. The precision of annual abundance estimators from the BSG were improved by an average of 43% over those based on the standard GSPE analysis, highlighting the very real costs of assuming a fixed (i.e., suboptimal) model structure. The population increased at a rate of 2.3%/year (95% CrI = 0.8–3.8%), and the increase was largely explained by a parallel increase in wildfire extent (i.e., high quality moose habitat). These results suggest that our approach could be used to increase estimator efficiency or decrease future survey costs without any modifications to the basic protocol. While modification of the GSPE software is possible, practitioners may find the BSG approach more convenient for quickly developing model structures for a particular application, thereby allowing them to extract more information from existing and future datasets.     

Female and male fitness of a sexually deceptive orchid with a narrow distribution area: from phenotypic traits to spatial distribution patterns

• The Orchidaceae family presents one of the most extravagant pollination mechanisms: deception. While many studies on reproductive success have been performed on food‐deception orchids, less have been performed on sexually deceptive orchids. Here, we focused on Ophrys balearica P. Delforge, an endemic orchid of the Balearic Islands, to study its reproductive ecology, the spatio‐temporal variation of its reproductive success and the individual (floral display and geospatial position) and population parameters (patch size, shape and density) that affect its reproductive success.
• We performed hand‐pollination experiments, along with the recording of floral display parameters and GPS position of over 1,100 individuals from seven populations in two consecutive years. We applied, for the first time, GIS tools to analyse the effects of individual’s position within the population on the reproductive success. Reproductive success was measured both in male (removed pollinia) and female (fruit set) fitness.
• The results confirm that this species is pollinator‐dependent and mostly allogamous, but also self‐compatible. This species showed high values for the cumulative inbreeding depression index and high pollen limitation. Male fitness was almost equal to female fitness between years and populations, and reproductive success exhibited huge spatio‐temporal variation.
• Although we did not find strong correlations between floral display and reproductive success, patches with low‐plant density and individuals in the external portion of the population showed significantly higher plant fitness. These findings must be considered in conservation actions for endangered orchid species, especially considering that most orchids are strongly dependent on pollinators for their species’ fitness.

     

Infusing considerations of trophic dependencies into species distribution modelling

Community ecology involves studying the interdependence of species with each other and their environment to predict their geographical distribution and abundance. Modern species distribution analyses characterise species‐environment dependency well, but offer only crude approximations of species interdependency. Typically, the dependency between focal species and other species is characterised using other species’ point occurrences as spatial covariates to constrain the focal species’ predicted range. This implicitly assumes that the strength of interdependency is homogeneous across space, which is not generally supported by analyses of species interactions. This discrepancy has an important bearing on the accuracy of inferences about habitat suitability for species. We introduce a framework that integrates principles from consumer–resource analyses, resource selection theory and species distribution modelling to enhance quantitative prediction of species geographical distributions. We show how to apply the framework using a case study of lynx and snowshoe hare interactions with each other and their environment. The analysis shows how the framework offers a spatially refined understanding of species distribution that is sensitive to nuances in biophysical attributes of the environment that determine the location and strength of species interactions.     

Geospatial insights into the controls of microbialite formation at Laguna Negra,Argentina

Microbialites provide a record of the interaction of microorganisms with their environment constituting a record of microbial life and environments through geologic time. Our capacity to interpret this record is limited by an incomplete understanding of the microbial, geochemical, and physical processes that influence microbialite formation and morphogenesis. The modern system Laguna Negra in Catamarca Province, Argentina contains microbialites in a zone of carbonate precipitation associated with physico-chemical gradients and variable microbial community structure, making it an ideal location to study how these processes interact to drive microbialite formation. In this study, we investigated the geospatial relationships between carbonate morphology, geochemistry, and microbial community at the macro- (decimeter) to mega- (meter) scale by combining high-resolution imagery with field observations. We mapped the distribution of carbonate morphologies and allochtonously-derived volcaniclasts and correlated these with sedimentary matrices and geochemical parameters. Our work shows that the macroscale distribution of different carbonate morphologies spatially correlates with microbial mat distributions—a result consistent with previous microscale observations. Specifically, microbialitic carbonate morphologies more commonly occur associated with microbial mats while abiotically derived carbonate morphologies were less commonly associated with microbial mats. Spatial variability in the size and abundance of mineralized structures was also observed, however, the processes controlling this variability remains unclear and likely represent a combination of microbial, geochemical, and physical processes. Likewise, the processes controlling the spatial distribution of microbial mats at Laguna Negra are also unresolved. Our results suggest that in addition to the physical drivers observed in other modern environments, variability in the spatial distribution of microbialites and other carbonate morphologies at the macro- to megascale can be controlled by microbial processes. Overall, this study provides insight into the interpretation of microbialite occurrence and distributions in the geologic record and highlights the utility of geospatial statistics to probe the controls of microbialite formation in other environments.     

Ecological factors affecting the fitness of the threatened orchid Anacamptis robusta (Orchidaceae): Habitat disturbance,interactions with a co‐flowering rewarding orchid and hybridization events

• The food‐deceptive species Anacamptis robusta is threatened in the Balearic Islands, and its habitat has recently been transformed through human disturbance. This study investigated how human disturbance affects the reproductive output of A. robusta and how its fitness is affected by competition with rewarding relatives, fungal infections and hybridization processes.
• To evaluate the impact of habitat loss on plant fitness, data on reproductive measures were obtained in two well‐conserved subpopulations and the unique disturbed subpopulation. Photo‐trapping cameras were installed to determine the floral visitation rate. All flowering individuals in 2019 were georeferenced using differential GPS to examine the influence of geospatial patterns on the reproductive success of A. robusta. In addition, hand‐pollination treatments were performed to evaluate the hybridization between A. coriophora and A. robusta and the origin of A. × albuferensis.
• The human‐disturbed subpopulation of A. robusta had a lower fruit set success than the subpopulations in well‐conserved areas. The presence of A. coriophora is negatively affecting the reproductive output of A. robusta. Moreover, A. robusta can only act as the pollen donor during hybridization.
• The complexity of the ecological system, which is enhanced by the strong pollinator dependence of the threatened species, must be considered when making conservation decisions. Although human disturbance directly affects plant population stability, other ecological issues must be considered, such as pollinator interaction, interspecific competition for pollinators, fungal infection and hybridization events.

     

A deep learning approach to conflating heterogeneous geospatial data for corn yield estimation: A case study of the US Corn Belt at the county level

Understanding large‐scale crop growth and its responses to climate change are critical for yield estimation and prediction, especially under the increased frequency of extreme climate and weather events. County‐level corn phenology varies spatially and interannually across the Corn Belt in the United States, where precipitation and heat stress presents a temporal pattern among growth phases (GPs) and vary interannually. In this study, we developed a long short‐term memory (LSTM) model that integrates heterogeneous crop phenology, meteorology, and remote sensing data to estimate county‐level corn yields. By conflating heterogeneous phenology‐based remote sensing and meteorological indices, the LSTM model accounted for 76% of yield variations across the Corn Belt, improved from 39% of yield variations explained by phenology‐based meteorological indices alone. The LSTM model outperformed least absolute shrinkage and selection operator (LASSO) regression and random forest (RF) approaches for end‐of‐the‐season yield estimation, as a result of its recurrent neural network structure that can incorporate cumulative and nonlinear relationships between corn yield and environmental factors. The results showed that the period from silking to dough was most critical for crop yield estimation. The LSTM model presented a robust yield estimation under extreme weather events in 2012, which reduced the root‐mean‐square error to 1.47 Mg/ha from 1.93 Mg/ha for LASSO and 2.43 Mg/ha for RF. The LSTM model has the capability to learn general patterns from high‐dimensional (spectral, spatial, and temporal) input features to achieve a robust county‐level crop yield estimation. This deep learning approach holds great promise for better understanding the global condition of crop growth based on publicly available remote sensing and meteorological data.