USDA Forest Service

Pacific Southwest Research Station

Pacific Southwest
Research Station

800 Buchanan Street
Albany, CA 94710-0011
(510) 883-8830
United States Department of Agriculture Forest Service. USDA logo which links to the department's national site. Forest Service logo which links to the agency's national site.

Research Topics Forest Management

Variable Density Thinning Study

second-growth forest on the Stanislaus-Tuolumne Experimental Forest
Typical contemporary condition of second-growth forest on the Stanislaus-Tuolumne Experimental Forest following logging in 1929 and over a century of fire exclusion.

Background and project overview

Forests of the Sierra Nevada were historically highly variable but generally considerably less dense than they are today because of frequent fire. In 1900, one noted early observer, Charles Sudworth, wrote that "forests of large, mature timber are rarely if ever dense; the single big trees, or groups of 3-6 trees stand far apart, forming a characteristically open forest". Numerous small gaps interspersed with the trees created habitat diversity with patches of shrubs and sun-loving forbs. These high light environments were also beneficial for trees such as pines that do not regenerate well in the shade. Other areas, particularly in the more productive lower slope positions, were often denser, providing habitat for wildlife species that prefer more shaded environments. This matrix of structures, along with light surface fuel loads, made forests relatively immune to stand-replacing crown fire.

The relative openness of forests was attributed to frequent fire, which many early foresters saw as a negative. It was believed that if fire could be kept out the forest could support many more trees. This became one of the main arguments for suppressing fire.

Once fire was removed from the system, understory vegetation gradually gave way to dense stands of young conifers - particularly white fir and incense cedar, which are more shade tolerant. Gaps began to fill with trees. Many of the larger most fire-resistant trees were cut for lumber. The forests we have today are generally far denser and more homogeneous than their historic counterparts. Historical Methods of Cutting study plots in the Stanislaus-Tuolumne Experimental Forest show that the contemporary forest is about 2.4x denser than old growth forest was in 1929. With the added density plus accumulated surface fuels, forests are also more vulnerable to stand-replacing wildfire and mortality from drought, bark beetles, and other stressors.

The standard silvicultural practice for reducing fire hazard often involves thinning from below to remove smaller trees (ladder fuels) that can propagate fire into the forest canopy, and spacing the crowns of the remaining trees relatively evenly. The effectiveness of such thinning for reducing the likelihood of stand-replacing wildfire is well established, but whether even crown spacing is the most efficient means of mitigating crown fire hazard in the long-term is not well understood. In addition, the relative lack of heterogeneity may reduce habitat suitability for wildlife species requiring a range of conditions, from closed canopy forest to openings and gaps. Natural regeneration of tree species such as sugar pine and ponderosa pine may also be limited without the presence of higher light environments such as those found in gaps.

Starting in 2009, the Pacific Southwest Research Station, partnering with the Stanislaus National Forest, set out to test some new approaches to restoring more resilient forest conditions. Among the objectives of the "Variable Density Thinning" study were to 1.) investigate whether a thinning prescription modeled after historical stand structures and designed to generate a high degree of structural variability would benefit a broader array of associated plant and animal species than a more standard even crown spacing treatment 2.) determine how overstory heterogeneity influences tree regeneration and maintains understory heterogeneity over time, 3.) evaluate whether leaving some trees in groups would sacrifice tree health and growth, and 4.) learn the degree to which heterogeneity can be generated with prescribed fire alone as well as the importance of fire in combination with thinning for promoting desirable ecological outcomes.

Map displays ares of the three different treatments (even thin, variable thin, and prescribed burn) and control units among the study units.
Map of the Variable Density Thinning Study units on the Stanislaus-Tuolumne Experimental Forest.

Study Area

The study is located in the Stanislaus-Tuolumne Experimental Forest, a 1700 acre area of the Stanislaus National Forest, near Pinecrest, CA, set aside for research purposes. The experimental forest was formally established in 1943 to represent high site quality mid-elevation (5200 and 6400 ft) mixed conifer forest, but scientists have been doing research in this area since shortly after the US Forest Service was established in 1905. While white fir is the most common tree species, the forest contains a high percentage of sugar pine compared with other forests of the Sierra Nevada. The 250 acre study area chosen for the 'Variable Density Thinning' study was railroad logged in the late 1920’s – this logging removed many of the trees >24” in size. However, due to the high productivity of the site, trees in excess of 50” in diameter can now be found.


Three thinning treatments (high variability "HighV" thin, low variability "LowV" thin, and unthinned control, all with or without prescribed fire were randomly assigned to units approximately 10 acres (4 ha) in size. Each of the six treatments was replicated four times, for a total of 24 units (Fig. 1). The HighV treatment units were thinned using a prescription designed to produce a highly variable structure similar to what historical forests once had. Historical data that informed the prescription were obtained from three "Methods of Cutting" plots installed in 1929 adjacent to the current study area. All trees >4" (10 cm) diameter in these historical plots were mapped and tree species and size measurements were collected prior to harvesting, providing useful reference information about the historical unlogged forest condition in this area (see 'Methods of Cutting' study for details).

The goal of the HighV thinning prescription was to produce a spatial structure, density, species composition, and size distribution consistent with the historical patterns once observed on this site. This included creating approximately one gap 0.1 to 0.5 ac (0.04 to 0.2 ha) in size every two acres (0.8 ha), which approximates the density of similarly-sized openings noted in these historical plots. The remainder of each unit was divided into groups of adjacent, similar size trees. Within groups, the best trees (defined as generally the largest and most vigorous) were retained, regardless of crown spacing. Rather than thinning each group to the same density, about a third of the groups were thinned more heavily, a third moderately, and a third lightly. No black oaks were cut but all conifers within the drip line of black oaks were removed where damage to the oak could be avoided. Additional details about the thinning prescription and how leave trees were marked are provided in Knapp et al. (2012).

The LowV thinning treatment was marked for cutting by selecting leave trees spaced approximately 0.5 crown widths from nearest neighbors.

Because of the current lack of pine compared with historical conditions, leave tree priority among conifers for both thinning treatments was sugar pine > ponderosa/Jeffrey pine > incense cedar > white fir.

Thinning was conducted between July and September 2011. Tree boles were processed into logs and other forest products at the landings. Small trees and tops of larger trees were chipped and removed as biomass. Designated burn units were treated with prescribed fire Nov. 11-18, 2013.

Data collection

A 98 foot (30 m) grid was established across the study area and data collection is referenced to this grid. Most vegetation data are collected along belt transects that vary in width, depending on the variable measured. Variables measured before, during, and after treatment implementation include:

  • Trees, seedlings, and saplings – mortality, establishment, growth
  • Canopy cover
  • Understory vegetation
  • Small mammals
  • Bird communities
  • Snow accumulation and melt out in relation to canopy structure
  • Soil moisture, soil temperature, and CO2 efflux

Most vegetation variables are currently being measured every two years, while others are monitored as funding becomes available.

Results - highlights

  • Both thinning treatments resulted in densities of >10” trees and species composition similar to what old-growth forests in this area historically contained.
  • The board foot volume removed to establish the HighV and LowV treatments averaged about 14,000 ft per acre and allowed the thinning to pay for itself. Had a 30” diameter cap been used, volume would not have differed between the two thinning strategies.
  • Prior to treatment, the study site had a high density of Northern flying squirrels (Glaucomys sabrinus), which are an important food source for raptors including spotted owls. While numbers caught in live traps declined in thinned units following treatment, the overall population size in the study area did not change, illustrating the potential benefits of habitat heterogeneity.
  • Thinning treatments suffered far less tree mortality during and after the 2012-2015 drought than the unthinned controls. Basal area (the cross sectional area of live tree stems) declined 23% between 2014 and 2018 in the unthinned controls, while the basal area did not change in the thinned units, with mortality balanced by tree growth.
  • Between 10-20% more snow accumulated in the thinned units compared with the unthinned controls in the 2013 and 2014 water years. Differences in snow melt out date among treatments were inconclusive, in part because both years were unusually warm and dry.
  • Many understory plant species are responding most favorably to the combination of either type of thinning plus prescribed fire. Some shrubs, including Ceanothus – an important browse for deer – show the largest increase in the HighV thinning plus prescribed fire treatment. Germination of Ceanothus seeds is stimulated by fire and the presence of gaps provides suitable high-light environments.

If you are interested in a tour, please contact principle investigator and science lead for the Stanislaus-Tuolumne Experimental Forest, Eric Knapp. Or, through a partnership effort with Penn State University, you can experience a 33-minute virtual reality tour of the study units with audio, or explore on your own with 360° photographs.


  • California Energy Commission (2010). Quantifying the effect of biomass removal treatments on habitat quality.
  • USDA Agriculture and Food Research Initiative (2011). Variable thinning using historical stand structure data to create fire-resilient forests and enhance ecosystem services in a changing climate. Grant number 11-03859.
  • PSW Emphasis Area request for proposals (internal funding)
    • Variable density thinning and prescribed fire effects on wildlife habitat (2014).
    • Effectiveness of variable density thinning and prescribed fire for promoting resistance to drought-induced bark beetle-caused tree mortality (2017).
    • Drought-induced bark beetle tree mortality in the Sierra Nevada: intermediate-term effects of thinning and prescribed fire (2018).
    • Evaluating the potential of virtual reality immersive experiences for improving natural resource decision making in fire prone forested ecosystems (2018).