Long-term tree population growth can predict woody encroachment patterns

Abstract

Recent increases in woody plant density in dryland ecosystems-or "woody encroachment"-around the world are often attributed to land-use changes such as increased livestock grazing and wildfire suppression or to global environmental trends (e.g., increasing atmospheric carbon dioxide). While such changes have undoubtedly impacted ecosystem structure and function, the evidence linking them to woody encroachment is mixed, and the underlying processes are not fully understood. To clarify the role of demographic processes in changing woody plant abundance, we conducted a meta-analysis of tree age structures from 29 woodland populations across the interior western United States, estimating per-capita tree establishment rates over the last several hundred years using demographic models. We found only limited evidence of increases in per-capita tree establishment rates following 19th-century Euro-American settlement. On the contrary, our results showed that observed age structures dominated by young trees, often cited as evidence of woody encroachment driven by anthropogenic processes, can be largely predicted by a null model based only on steady, multiplicative tree population growth. Moreover, we demonstrated that tree establishment rates in the last century have mostly declined rather than increased, and they are currently at their lowest rates since at least 1600 CE. Our results suggest that a large part of modern increases in woodland tree establishment and density may in fact be a result of long-term population increases, and failing to consider the demographic processes underlying population growth can lead to an overestimation of settlement effects on stand structure.

Keywords: demography; null model; pinyon–juniper woodlands; range dynamics; woody encroachment.

Fig. 1.

Increasing establishment under a constant establishment rate. Because new establishment ($B_t$) is the product of both population size ($N_{t-1}$) and per-capita establishment rate ($b_t$; $B_t=b_t{N}_{t-1})$, in a growing population, the amount of new establishment will increase over time even if the underlying establishment rate is constant. One could view the rapid increase in establishment and population size as evidence of a change point and link the increase in establishment to an external driver. However, analysis of the underlying demographic mechanisms will find that no change in per-capita establishment rates (orange) has occurred. Increased establishment is instead driven by the reproductive potential of larger populations. BP indicates “before present.” (Illustration by Alex Boersma).

Fig. 2.

Tree establishment rates since 1600 CE. (A) Map of sites in which age structure data exist (points) overlayed on the range pinyon–juniper in the western United States. (B) Proportion of total observed establishment occurring in each 20-y interval averaged across 29 pre-1800 populations and cumulative observed establishment in each population broken down by species. Species acronyms in legend are mixed stands dominated by Pinus monophylla, with a minor Juniperus osteosperma component (PiMo-JuOs), Juniperus occidentalis (JuOc), Pinus monophylla (PiMo), Pinus edulis (PiEd), and Juniperus osteosperma (JuOs). (C) Per-capita establishment rates for each of the 29 pre-1800 populations (points), posterior median estimate for 20-y average (blue) and 95% credible interval (gray), and the best fit segmented regression (dashed orange). The entire timeseries dataset is shown in SI Appendix, Fig. S2.

Fig. 3.

Predictability of tree establishment patterns. (A) Bars representing average expected establishment in 20-y intervals across all populations expected from changing per-capita establishment rates (orange) and multiplicative population growth with a constant establishment rate (blue). (B) Proportion of variance in establishment explained by constant establishment rates in each population, colored by species. (C) Ratio of establishment ($\frac{\mathit{Constant}}{\mathit{Changing}}$) between each mechanism for each population in each 20-y interval. The solid black line indicates posterior estimate and 95% CI for the median. Values of 1 (dashed black line) indicate that the amount of establishment is the same under either mechanism, whereas values below 1 indicate that changing establishment rates would be expected to lead to greater establishment.

Fig. 4.

Impacts of survival rates on estimated per-capita establishment rates. (A) Lines representing individual posterior draws for survival rates for a single population illustrating wide variability. (B) Lines representing individual posterior draws for establishment rates for the same population illustrating well-constrained estimates. (C) Correlation of survival and establishment rate posterior draws in all populations illustrating low correlation.