Skip to main content

Data from: Tree growth responses to extreme drought after mechanical thinning and prescribed fire in a Sierra Nevada mixed-conifer forest, USA

Cite this dataset

Zald, Harold et al. (2022). Data from: Tree growth responses to extreme drought after mechanical thinning and prescribed fire in a Sierra Nevada mixed-conifer forest, USA [Dataset]. Dryad.


An estimated 128 M trees died during the 2012-2016 California drought, largely in the southern Sierra Nevada Range. Prescribed burning and mechanical thinning are widely used to reduce fuels and restore ecosystem properties, but it is unclear if these treatments improve tree growth and vigor during extreme drought. This study examined tree growth responses after thinning, prescribed burning, and extreme drought at the Teakettle Experimental Forest, a historically frequent fire mixed-conifer forest in the southern Sierra Nevada of California, USA. Mechanical thinning (no thin, understory thin, and overstory thin) and prescribed burning (unburned, fall burning) were implemented in 2000-2001. Using annual growth data from increment cores, over 10,000 mapped and measured trees, and lidar-derived metrics of solar radiation and topographic wetness, we had two primary questions. First, what were the growth responses to thinning and prescribed burning treatments, and did these responses persist during the 2012-2016 drought? Second, what tree-level attributes and environmental conditions influenced growth responses to treatments and drought?

Thinning increased residual tree growth and that response persisted through extreme drought 10 -15 years after treatments. Growth responses were higher in overstory versus understory thinning, with differences between thinning types more pronounced during drought. Species-specific growth responses were strongest with overstory thinning, with sugar pine (Pinus lambertiana) and incense-cedar (Calocedrus decurrens) having higher growth responses compared to white fir (Abies concolor) and Jeffery pine (Pinus jeffreyi). For individual trees, factors associated with higher growth responses were declining pretreatment growth trend, smaller tree size, and post-treatment low neighborhood basal area. Growth responses were initially not influenced by topography, but topographic wetness became important during extreme drought. Mechanical thinning resulted in durable increases in residual tree growth rates during extreme drought over a decade after thinning occurred, indicating treatment longevity in mitigating drought stress. In contrast, tree growth did not improve after prescribed burning, likely due to fire effects that reduced surface fuels, but had little effect on reducing tree density. Thinning treatments promoted durable growth responses, but focusing on stand-level metrics may ignore important tree-level attributes such as localized competition and topography associated with higher water availability. Mechanical thinning was effective at improving growth in trees that had been experiencing declining growth trends, but was less effective in improving growth responses in large old higher ecological importance.


See methods section of Zald et al. 2022. Tree growth responses to extreme drought after mechanical thinning and prescribed fire in a Sierra Nevada mixed-conifer forest, USA. Forest Ecology and Management.

Usage notes


RData file of data used for analysis in Zald et al. 2022. RData file contains three dataframes (tk.rwl, tk.lookup, and tk.comp.topo). tk.rwl contains raw ring width data for all sampled trees. tk.lookup contains tree attribute data for all sampled trees. tk.comp.topo contains local neighborhood basal area values, potential solar radiation values, and topographic wetness values for censused trees.


R Markdown html uses data frames in TEF_Dendro.Rdata to conducting the following data processing and analyses steps: 1. Plot sampled trees in relation to all live tree diameters and local (10 m radius) neighborhood basal area within each treatment combination. 2. Plot relationship between diameter and age by species for sampled trees. 3. Calculate tree growth metrics. 4. Plot annual growth trends over time by treatment combination. 5. Stand level growth responses using linear mixed effects models, with diagnostics, estimated marginal means and post-hoc comparison tests. 6. Tree level growth responses using random forest with multi-stage model selection.


Joint Fire Science Program, Award: 15-1-07-6

California Department of Forestry and Fire Protection, Award: 8GG14803