Semiarid grasslands and extreme precipitation events: do experimental results scale to the landscape?

Abstract

The frequency and magnitude of deluges (extremely large rain events) are increasing globally as the atmosphere warms. Small-scale experiments suggest that semiarid grasslands are particularly sensitive to both the timing and size of deluge events. However, the assumption that plot-scale results can be extrapolated across landscapes with variable soil textures, plant communities, and grazing regimes has seldom been tested, despite being key to forecasting regional consequences of precipitation extremes. We used precipitation data from an extensive rain gauge network to identify natural deluges (mean size = 60 ± 31 mm, 1984-2012) that occurred across a ~60 km 2 heterogeneous native shortgrass steppe landscape in Colorado. We then related spatial variation in deluge precipitation to post-deluge responses in canopy greenness (normalized difference vegetation index, NDVI) via satellite imagery. Consistent with results from experiments, this semiarid grassland was most sensitive to mid-growing season deluges, and post-deluge canopy greenness usually increased linearly (67% of the time) with increasing deluge size. This suggests that aboveground productivity in these semiarid systems will likely increase, rather than asymptote, with forecasted increases in deluge size. Importantly, differences in grazing regime did not significantly alter deluge responses, indicating that these patterns are robust to this widespread management practice.

Time-Series Characterization of Grassland Biomass Intensity to Examine Management Change at Two Sites in Germany Using 25 Years of Remote Sensing Imagery

“…S1: Source LULC polygon classes, levels, numbers and total area for Orbroich; Table S2: Source grassland polygon statistics for Wahnbachtal; Table S3: Year, season and number of images used in analysis for Orbroich and Wahnbachtal sites; Table S4: Correlation of various vegetation indices (Vis) considered for the study; Table S5: Evaluation of cluster separability of selected VIs using transformed divergence; Parcel classes used for grassland analysis are in bolded black font; Table S6: Comparison of cluster assignments using NDVI, MSI, and TVI vegetation indices for Orbroich and Wahnbachtal sites showing agreement between methods in grey shaded cells. References [9,10,[31][32][33] are cited in the Supplementary Materials.…”

“…Figure S5: Two-stage method of image normalization showing the master base image (dark orange) and base images for each year (light orange), and within-year images normalized to the base image for that year (each row) for the Orbroich study site; Figure S6: Grassland clusters (left) with temporal adjustment to account for annual factors affecting entire areas (center) resulting in final annual NDVI values for each cluster (right) in Orbroich (top) and Wahnbachtal (bottom); Figure S7: Comparison of clustering using all available images across each year (A) and only the spring base image (B) for the Orbroich site; Figure S8: Statistical analysis of cluster; Figure S9: Cluster variability for Orbroich site using the standard deviation of individual grassland parcel NDVI values for each year; Figure S10: Cluster variability for Wahnbachtal site using the standard deviation of individual grassland parcel NDVI values for each year; TableS1: Source LULC polygon classes, levels, numbers and total area for Orbroich; TableS2: Source grassland polygon statistics for Wahnbachtal; TableS3: Year, season and number of images used in analysis for Orbroich and Wahnbachtal sites; TableS4: Correlation of various vegetation indices (Vis) considered for the study; TableS5: Evaluation of cluster separability of selected VIs using transformed divergence; Parcel classes used for grassland analysis are in bolded black font; TableS6: Comparison of cluster assignments using NDVI, MSI, and TVI vegetation indices for Orbroich and Wahnbachtal sites showing agreement between methods in grey shaded cells. References[9,10,[31][32][33] are cited in the Supplementary Materials.…”

Compound hydroclimatic extremes in a semi‐arid grassland: Drought, deluge, and the carbon cycle

“…To assess plant growth and phenology, weekly plot canopy greenness was estimated using repeat digital photography (adapted from Post & Knapp, 2020, 2021). Previous research has demonstrated that plot‐level responses in canopy greenness scale well to landscape‐level measurements (e.g., satellite‐derived greenness via normalized difference vegetation index, or NDVI; Post et al, 2021). For each photograph, a digital camera was positioned directly over a marked region of interest (50 × 50 cm frame in the corner of each plot) at a 90° angle relative to the ground surface.…”