Fig. 5 | Nature Communications

Fig. 5

From: Sampling bias overestimates climate change impacts on forest growth in the southwestern United States

Fig. 5

Regional ring-width index (RWI) chronologies for the U.S. Southwest (all samples south of 38°N) based on the forest inventory sample (FIA, in red) vs. targeted sample (ITRDB, in black). The gray envelopes indicate variability among 100 iterations of randomly resampling available ITRDB ring-width time series to match the sample replication of the FIA chronology. Boxplots show projected mean growth for 2010–2039, 2040–2069, and 2070–2099 under the “business-as-usual” scenario RCP8.5 across 15 different Atmosphere-Ocean Global Circulation Models (GCM) from the CMIP5 ensemble. Bold horizontal lines denote the median of the projections, the boxes denote the interquartile range, and whiskers extend to the most extreme future RWI values arising from GCM-projected future climate. Asterisks show the significance level of differences in projected future tree growth based upon the two datasets (two-sided t test; *p < 0.05; ***p < 0.001). Sample replication, i.e., the total number of ring-width time series, is indicated as counts (right-hand y axis) in gray (ITRDB) and red (FIA) shaded areas below the chronologies (right-hand y axis)

Back to article page