Jonas Trepel

@jonastrepel.bsky.social

2013 Followers 508 Following

Postdoctoral researcher at Aarhus University. Broadly interested in conservation biology, rewilding, megafauna ecology, global change & biodiversity.

Figure showing that protection status does not moderate the relationship between changes in green plant biomass and global change drivers. (A) Spatial distribution of strictly protected areas (IUCN category Ia, Ib and II) and paired control areas in similar environmental conditions and of the same size and shape; (B) interaction of global change drivers and protection status. The x-axis is restricted to the 2.5–97.5th percentile of the respective variable. The units for the predictors are as follows: fire frequency is given in number of fires between 2001 and 2024, human modification index (HMI) change is in percent of land modified, mean annual temperature (MAT) trend and max. annual temperature trend is °C year−1, N deposition is in kg km−2 year−1 and precipitation trend is mm year−1. Shading represents the 95% confidence intervals. All variables have been scaled and centered for modeling but are plotted on their original scale to ease interpretation. The models further contained MAT and mean annual precipitation (MAP) to account for the environmental context (MAT: ß = 0.33 [95% CI: −0.24, 0.91]; MAP: ß = −0.01 [−0.37, 0.36]. Variables were considered significant when the 95% confidence interval of the coefficient estimate did not overlap zero.
Figure showing positive association between N deposition and increases in green plant biomass across biomes. Models for changes in EVI as a function of N deposition were fitted for a variety of subsets separately, including Köppen–Geiger climatic regions (A)–(B) and aggregated biomes based on the functional biomes of Higgins et al. (2016) (C)–(D). (A) Coefficient estimates for each Köppen–Geiger climatic region separately. Error bars indicate the 95% confidence interval; (B) spatial distribution of climatic regions; (C) coefficient estimates for aggregated biomes separately; (D) spatial distribution of aggregated biomes. All variables have been scaled and centered for modeling. Variables were considered significant when the 95% confidence interval of the coefficient estimate did not overlap zero. A positive coefficient estimate indicates a positive relationship between the respective predictor variable and the response.
Fig. 1 of the manuscript: Global increases in green plant biomass are strongly linked to atmospheric N deposition. (A) Spatial distribution of trends calculated with an autoregressive model that accounts for temporal autocorrelation; EVI was scaled by 10 000. (B) Model estimates from a generalized linear mixed effects model which accounts for spatial autocorrelation using spatial random fields. The x-axis is restricted to the 2.5–97.5th percentile of the respective variable; see the Supporting information for the full range and raw data points. A solid line indicates a statistically significant relationship, a dashed line a non-significant relationship. The units for the predictors are as follows: fire frequency is given in number of fires between 2001 and 2024; human modification index (HMI) change is the difference in percent of land modified between 2020 and 2000; mean annual temperature (MAT) is in °C and mean annual precipitation in mm, MAT trend and max. annual temperature trend is °C year−1 (from 1951 to 2024), N deposition is in kg km−2 year−1 and precipitation trend is mm year−1 (from 1951 to 2024). All variables have been scaled and centered for modeling but are plotted on their original scale to ease interpretation.
The field team crawling through the thicket inside an exclosure
A buffalo, disturbed while grazing and looking mildly displeased
Effects of large herbivores on plant taxonomic diversity, dominance, functional diversity and vegetation structure. Estimates from intercept only GLMMS for plant community responses on plot (n = 108 pairwise comparisons) and site scale (n = 6 pairwise comparisons). Plant functional distance serves as a proxy for functional redundancy, with a lower distance representing a higher redundancy. Orange points right of the dashed line represent a significant increase of the variable in the presence of herbivores and orange points left of the dashed line represent a significant decrease in the presence of herbivores. Points in the background are the raw effect sizes, coloured by their value. Error bars represent the 95% confidence interval.
Study area, exclosure locations and plot design. (a) Study area. The shaded polygons represent the protected areas we worked in (Addo Elephant National Park [AENP], Phongolo Nature Reserve [PNR], Kruger National Park [KNP]). The inset maps show the exclosures and the black dots correspond to the true plot locations. The Nyathi exclosures extend approximately 200 m at their longest site, while Jack's exclosure reaches about 2200 m. The PNR border fence is approximately 4400 m long and 20 m wide, and therefore appears as a line on the map. The black dots indicate clusters of three plots inside and three paired plots outside the fence. The Nkuhlu exclosure spans approximately 1350 m, the Satara exclosure about 550 m and the N'waswitshumbe exclosure approximately 2000 m; (b) plot layout. We identified all herbaceous plants on small plots (0.5 × 2 m), nested in bigger plots (2.5 × 10) on which all woody plants were identified. The LiDAR scanner was placed in the centre of the larger plot.
Effects of large herbivores on plant taxonomic diversity, dominance, functional diversity and vegetation structure. Estimates from intercept only GLMMS for plant community responses on plot (n = 108 pairwise comparisons) and site scale (n = 6 pairwise comparisons). Plant functional distance serves as a proxy for functional redundancy, with a lower distance representing a higher redundancy. Orange points right of the dashed line represent a significant increase of the variable in the presence of herbivores and orange points left of the dashed line represent a significant decrease in the presence of herbivores. Points in the background are the raw effect sizes, coloured by their value. Error bars represent the 95% confidence interval.
Study area, exclosure locations and plot design. (a) Study area. The shaded polygons represent the protected areas we worked in (Addo Elephant National Park [AENP], Phongolo Nature Reserve [PNR], Kruger National Park [KNP]). The inset maps show the exclosures and the black dots correspond to the true plot locations. The Nyathi exclosures extend approximately 200 m at their longest site, while Jack's exclosure reaches about 2200 m. The PNR border fence is approximately 4400 m long and 20 m wide, and therefore appears as a line on the map. The black dots indicate clusters of three plots inside and three paired plots outside the fence. The Nkuhlu exclosure spans approximately 1350 m, the Satara exclosure about 550 m and the N'waswitshumbe exclosure approximately 2000 m; (b) plot layout. We identified all herbaceous plants on small plots (0.5 × 2 m), nested in bigger plots (2.5 × 10) on which all woody plants were identified. The LiDAR scanner was placed in the centre of the larger plot.
Screenshot of the paper title page, showing the full authorlist: Jonas Trepel, Joe Atkinson, Elizabeth le Roux, Andrew Abraham, Marge Aucamp, Michelle Greve, Marilize Greyling, Jesse Kalwij, Steven Khosa, Lukas Lindenthal, Caroline Makofance, Londiwe Mokoena, Anika Oosthuizen, Bent Rech, Erick Lundgren, Jens-Christian Svenning, Robert Buitenwerf
Graphical abstract for our new paper, showing sketches of two reserves, one with and one without large herbivores. The one with large herbivores has higher herbaceous species richness and functional redundancy.
Beautiful Sunset in Kaingo Private Game Reserve, one of our fieldsites
The figure shows the predicted plant nitrogen concentration in a radius of 2500m around the enclosures/feeding sites, clearly highlighting higher plant nutrient concentrations closer to the enclosures.
Figure 2 from Rech et al. (2025) shows habitat selection estimates for cattle, horses, and buffalo, separated by feeding status and season.
Name tag and brochure of the Waterberg Research Convention
A white rhino in the savanna