The dataset "data_fig2.csv" contains the information used to create Fig. 2 in the manuscript. The columns in this dataset are: "classe1": height class in meters "disturb": type of disturbance (i.e., fire, wind, or logging) "prop.before": the return probability before the disturbance "prop.after": the return probability after the disturbance "diff.prop": the difference of the return probabilities listed above The dataset "data_fig3.csv" contains the information used to fit the GAM model. The inference from this model is depicted in Fig. 3 of the manuscript. The columns in this dataset are: "classe1": height class in meters "tot": number of incoming returns "n": number of returns "xcat.std": standardized x spatial coordinate "ycat.std": standardized y spatial coordinate "ph": standardized ph of the soil "sand": standardized proportion of sand in the soil "l1.cec": standardized log(cation exchange capacity + 1) "l1.soc": standardized log(soil organic carbon + 1) "l1.wtd": standardized log(-water table depth + 1) "l1.slope1": standardized log(terrain slope + 1) "l1.elev": standardized log(terrain elevation + 1) "l1.road": standardized log(distance to nearest road + 1)