Deep learning to extract the meteorological by-catch of wildlife cameras

Microclimate-proximal climatic variation at scales of metres and minutes-can exacerbate or mitigate the impacts of climate change on biodiversity. However, most microclimate studies are temperature centric, and do not consider meteorological factors such as sunshine, hail and snow. Meanwhile, remote cameras have become a primary tool to monitor wild plants and animals, even at micro-scales, and deep learning tools rapidly convert images into ecological data. However, deep learning applications for wildlife imagery have focused exclusively on living subjects. Here, we identify an overlooked opportunity to extract latent, ecologically relevant meteorological information. We produce an annotated image dataset of micrometeorological conditions across 49 wildlife cameras in South Africa's Maloti-Drakensberg and the Swiss Alps. We train ensemble deep learning models to classify conditions as overcast, sunshine, hail or snow. We achieve 91.7% accuracy on test cameras not seen during training. Furthermore, we show how effective accuracy is raised to 96% by disregarding 14.1% of classifications where ensemble member models did not reach a consensus. For two-class weather classification (overcast vs. sunshine) in a novel location in Svalbard, Norway, we achieve 79.3% accuracy (93.9% consensus accuracy), outperforming a benchmark model from the computer vision literature (75.5% accuracy). Our model rapidly classifies sunshine, snow and hail in almost 2 million unlabelled images. Resulting micrometeorological data illustrated common seasonal patterns of summer hailstorms and autumn snowfalls across mountains in the northern and southern hemispheres. However, daily patterns of sunshine and shade diverged between sites, impacting daily temperature cycles. Crucially, we leverage micrometeorological data to demonstrate that (1) experimental warming using open-top chambers shortens early snow events in autumn, and (2) image-derived sunshine marginally outperforms sensor-derived temperature when predicting bumblebee foraging. These methods generate novel micrometeorological variables in synchrony with biological recordings, enabling new insights from an increasingly global network of wildlife cameras.

holobiont reveals a microbiome resilient to future climate changes

The plant microbiome supports plant growth, fitness, and resistance against climate change. Trifolium pratense (red clover), an important forage legume crop, positively contributes to ecosystem sustainability. However, T. pratense is known to have limited adaptive ability toward climate change. Here, the T. pratense microbiomes (including both bacteria and fungi) of the rhizosphere and the root, shoot, and flower endospheres were comparatively examined using metabarcoding in a field located in Central Germany that mimics the climate conditions projected for the next 50-70 years in comparison with the current climate conditions. Additionally, the ecological functions and metabolic genes of the microbial communities colonizing each plant compartment were predicted using FUNGuild, FAPROTAX, and Tax4Fun annotation tools. Our results showed that the individual plant compartments were colonized by specific microbes. The bacterial and fungal community compositions of the belowground plant compartments did not vary under future climate conditions. However, future climate conditions slightly altered the relative abundances of specific fungal classes of the aboveground compartments. We predicted several microbial functional genes of the T. pratense microbiome involved in plant growth processes, such as biofertilization (nitrogen fixation, phosphorus solubilization, and siderophore biosynthesis) and biostimulation (phytohormone and auxin production). Our findings indicated that T. pratense microbiomes show a degree of resilience to future climate changes. Additionally, microbes inhabiting T. pratense may not only contribute to plant growth promotion but also to ecosystem sustainability.

Viral infection can reduce the net nitrogen inputs of legume break crops and cover crops

Legumes are used in crop rotations by both large-scale and smallholder farmers alike to increase soil fertility, especially before high-nitrogen-demanding crops such as corn (maize). Legume crop residues and green manures are rich in nitrogen due to mutualistic rhizobia, bacteria that live in their roots and convert atmospheric nitrogen into a biologically available form. Growers can obtain recommendations from local extension offices about how much less inorganic nitrogen fertilizer needs to be added to a subsequent crop following different legume break crops for the predominant soil type (the nitrogen fertilizer replacement value, or NFRV). Due to the intimate relationship between legumes and rhizobia, conditions that affect plant health can also affect the rhizobia and how much nitrogen they provide. We use a combination of empirical data and previously published values to estimate reductions in nitrogen inputs under outbreaks of plant viruses of varying severity. We also use historical fertilizer prices to examine the economic impacts of this lost fertilizer for farmers. We find that fertilizer losses are greatest for crops that fix large amounts of nitrogen, such as clover and alfalfa as opposed to common bean. The economic impact on farmers is controlled by the proportion of plants with viral infections and the price of synthetic fertilizer. In a year of high disease prevalence, attention is normally focused on the yield of the diseased crops. We suggest that farmers growing legumes as break crops should be concerned about yields of subsequent crops as well. Viral diseases can be difficult to diagnose in the field, so the easiest way for farmers to prevent unexpected yield losses in subsequent crops is to test their soil when it is feasible to do so.

(red clover)

Re-growth of fodder plants after grazing and mowing drives the profitability of their cultivation and is therefore an important target trait for plant breeding and agricultural engineering. However, for some fodder plants little is known about their re-growth dynamics in response to grazing or mowing. We analysed the native response of plant architecture, leaf morphology and growth performance to experimental cutting in wild Trifolium pratense L. (red clover) plants. A total of 150 potted clover plants were established under controlled field conditions, and half of the plants were cut to 5 cm 3 months after sowing. Each plant was measured every week for 5 months. The cut and subsequently re-grown plants carried fewer main branches (-20%), as well as fewer (-13%) and smaller (-32%) leaves than the control plants. However, the cut plants produced an average of 17% more accumulated leaf area (cut + re-grown leaf area) than the control plants. This discrepancy was explained by variation in the growth strategy of the plants, where the cut plants invariably expressed a second growth phase, while almost half of the untreated plants did not. Our results suggest that cutting acted as an artificial trigger initiating a second growth phase in the cut plants and thereby contributed to yield increase. Exploiting this mechanism may set new goals for breeding and optimisation of the mowing regime.