This is the ReadMe file for the source code and datasets contained in the publication: Holdo, R.M. and Onderdonk. D.A. Savanna tree abundance and spatial patterns are strongly associated with river networks in Serengeti National Park, Tanzania. Landscape Ecology 2022.

Contact authors for all files: Ricardo Holdo (rholdo@uga.edu)

File Final_Analysis.R This file replicates all analyses and figures in the paper The code draws from the following data files:

File "Final_30m_data.csv" Raw tree cover data at 30-m resolution (after cubic interpolation). Columns are as follows: ID: Plot ID TC: Tree cover fraction X: Eastings (Datum: WGS84, Projection: UTM 36S) Y: Northings driver: Distance to river (km)

File "Dist_to_riv_spatial_regressions_poly.csv" Polynomial fits for regressions of TC vs. distance to river. Columns are as follows: ID: Plot ID AIC.null: AIC values for null (intercept) model AIC.lin: AIC values for linear model AIC.quad: AIC values for quadratic model AIC.cub: AIC values for cubic model

File "Distance_to_riv_cutoff.csv" Breakpoint for first segment of segmented regression of TC vs driver. Columns are as follows: ID: Plot ID psi: Breakpoint distance (km) Slope: slope of first segment (1/km), based on gls regression of log TC vs. driver p.val: P-value of gls regression

File "Aggregated_lacunarity_Data.csv" Lacunarity data at three spatial scales. Columns are as follows: ID: Plot ID Dbin: Distance bin (m). Value represents midpoint in 50-m intervals from river. Lac10: Normalized lacunarity at 10-m scale Lac25: Normalized lacunarity at 25-m scale Lac50: Normalized lacunarity at 50-m scale dr: Distance to river (km) dr2: dr ^ 2 dr3: dr ^ 3