Relating talus characteristics to habitat use by vulnerable alpine mammals in a changing climate

Ana Torres Ferreira

doi:10.7273/000007484

Structurally complex habitats play a critical role in the survival of many wildlife species by providing food, thermal refuges, and protection from predators. However, the same complex properties that make these habitats suitable for wildlife also make them challenging to characterize and assess their ecological value. Traditional methods for characterizing these habitats are labor-intensive and lack the fine-scale resolution needed for accurate habitat assessment. Therefore, we developed a novel workflow using handheld photogrammetry to systematically map and measure fine-scale habitat characteristics in broken-rock patches, evaluate its ecological applications, and identify key habitat features relevant to rock-dwellers. To test this workflow, we established 10 m x 10 m plots at 19 talus patches within the North Cascades National Park Service Complex in Washington and used a structured approach to capture overlapping images with an iPad Pro. We processed images in Agisoft Metashape, using structure-from-motion (SfM), to create georeferenced digital elevation models (DEMs) and orthomosaics. We used the Python package ‘segmenteverygrain’ to semi-automatically segment the orthomosaic rasters into individual rocks and enable us to quantify size and shape. For classifying areal cover, we created training samples and used a support vector machine algorithm to classify rasters into three categories (rock, moss, and vegetation). Using the DEMs, we characterized surface roughness at a 15-cm resolution. SfM photogrammetry successfully generated high-resolution orthomosaics and DEMs for all surveyed plots, with a ground sampling distance ≤ 1.17 mm. Image segmentation and supervised classification provided detailed quantitative data, revealing variations in rock size, shape, cover types, and surface roughness characteristics. This workflow provides a scalable and adaptable tool for fine-scale habitat characterization of structurally complex systems, providing relevant information for species and habitat monitoring. One iconic obligate dweller of broken-rock habitats is the American pika (Ochotona princeps). Unlike many alpine species, pikas do not hibernate and rely on caching vegetation in haypiles to survive during the winter when vegetation is scarce. Although many studies have evaluated the influence of patch-scale features on habitats selected by these obligate rock-dwellers, little is known about how fine-scale rock characteristics affect pika haypile site selection, because of challenges in collecting systematic, high-resolution data in talus patches. Therefore, we used the handheld photogrammetry method we developed to measure rock size and shape, surface roughness, and vegetation cover in 3 x 3 m plots around pika haypiles and 2 associated control plots at 40 sites within the North Cascades National Park and Mt Baker-Snoqualmie National Forest in Washington. We compared models predicting haypile site selection from these metrics and 2 patch-level metrics, proximity to patch edge and location within the talus patch. The top model predicting haypile site selection included rock size (maximum area), shape (flatness), surface roughness, and vegetation cover. Pikas selected haypile plots with larger and less flat rocks and rougher plots with less vegetation than control plots. The model with patch-level variables alone did not perform better than the null model. Using a portable and easy-to-apply photogrammetry method, we demonstrated that pikas respond to fine-scale rock characteristics when selecting haypile sites. Habitat degradation, driven by natural and anthropogenic factors, may pose a significant threat to American pikas, and our study emphasizes the need for conservation efforts that preserve both occupied talus patches and their structural integrity to support pika persistence in a rapidly changing world. Not only does broken-rock land cover provide suitable conditions for food storage, it also provides critical microhabitat for small vertebrates that buffers them from extreme temperatures. Although many studies have evaluated the temperature difference between the subsurface (i.e., crevices) and the surface of the talus patch, they have mainly focused on patch-scale variables and have often been limited to short-term observations. This study examined the temperature patterns of 12 talus patches in the North Cascades National Park Service Complex in Washington, over a 10-year period (2015-2024) across seasons, years, and diurnal periods between crevice and surface temperatures. We used our handheld photogrammetry method to characterize rock size, rock shape, vegetation and moss cover, and southwest-facing rock surface area to examine their influence on crevice and surface temperatures. We found that summer surface temperatures increased by 0.37°C on average, per year, while winter surface temperatures remained stable. Crevices buffered high midday and evening temperatures in summer and retained warmth during the cold winter and at night, with average temperatures 5°C cooler than the surface in summer and 0.45°C warmer in winter. Surface and crevice temperatures were influenced by solar radiation, southwest-facing rock surface area, vegetation and moss cover, rock circularity, and elevation. Our findings highlight the value of collecting precise measurements of plot-level rock characteristics using handheld photogrammetry to better understand the thermal value of rocky habitats for wildlife. Crevices played a crucial role in buffering temperature extremes, underscoring their importance for alpine species, and as climate change accelerates warming, these microhabitat refugia will likely become even more essential for species survival.

Relating talus characteristics to habitat use by vulnerable alpine mammals in a changing climate

Files and links (1)

Abstract

Metrics

Details