Study team

Plan Roadmap

Goal framework

Core question

Goals forum

Wellbeing concerns

Strategy portfolio


Social portfolio

Silviculture portfolio

Operations portfolio

Stand-scale studies

Secondary questions

Statistical models


Work completed





1: Interim guidance

2: Strategy rationale

3: Funding

4: Workloads

Study Plan

Large-Scale Integrated Management Experiment
on the Olympic Experimental State Forest
(Short Name: T31 Experiment)

Version: October 20, 2017


Study Team

Name Affiliation Role Expertise
Bernard Bormann UW Study PI Sustainability, applied forest ecology
Teodora Minkova WADNR Study PI Watershed management
Bill Wells WADNR Study PI Operations, silviculture
Marc Miller UW Study PI Sustainability, applied social science
Brian Harvey UW Scientist Forest ecology and management
Dan Donato WADNR Scientist Forest ecology and management
David Butman UW Scientist Biogeochemistry
Dede Olson PNW Scientist Riparian ecology
Drew Rosenbalm WADNR Manager Silviculture, operations
Frank Hanson UW Outreach Education, tribal liaison
Gordon Reeves PNW Scientist Aquatic systems
Greg Ettl UW Scientist Silviculture
Keven Bennett UW Support GIS, web
Kyle Martens WADNR Scientist Fish ecology
Paul Anderson PNW Scientist Silviculture
Peter Kiffney NOAA Scientist Aquatic systems
Richard Bigley WADNR Scientist Forest ecology and management
Ryan Bellmore PNW Scientist Aquatic ecology
Sandor Toth UW Scientist Operations research
Scott McLeod WADNR Scientist Silviculture
Warren Devine WADNR Support Database and analyses
Woodam Chung OSU Scientist Operations research
As yet unconfirmed
Connie Harrington PNW Scientist Silviculture
Derek Churchill UW Scientist Forest ecology

We anticipate additions to the study team as needs and interests emerge.


Study Plan Roadmap


This plan describes a major management experiment on Washington Department of Natural Resources (DNR) Trustlands on the outer Olympic Peninsula. This experiment will take place on the Olympic Experimental State Forest (OESF) with leadership from DNR and the Olympic Natural Resources Center (ONRC).  The study plan development will be scientifically-vetted and respond to comments and suggestions from a wide array of partners and stakeholders. At its simplest, the study compares a range of management strategies that variably integrate environmental and community wellbeing goals at a landscape scale through normal operations of DNR managers.  This study is an example of science-based adaptive management where rigorous science is applied to address the most important management questions in ways that link to specific future decisions[Bormann et al. 2007].   The study opens the door to innovations in practice (within legal sidebars) that might be incorporated into standard practice in the future DNR land management. With the exception of its adaptive management guidance and legal sidebars, the study is partly decoupled from the current specifics of the DNR land management plan on purpose.  Management plans, including conservation strategies and sustainable harvest levels, have changed and will continue to change in part because of evolving societal needs and perspectives.  Accordingly, we begin with a review of the goal framework to assure that the study provides the greatest value over the long run. 

Collaborative approach to the study plan

This study is complex, requiring collaboration among a wide range of scientific disciplines and professional experts, and rapid feedback from stakeholders. Communication within and between these groups is not easy.  We are trying to develop a common lexicon and provide a glossary to improve clarity as much as possible. 

The HTML format here is designed to allow participants to jump to different parts of the plan easily. This initial version, designed to start the process, is preliminary and needs many inputs before becoming a coherent first draft.   The vertical HTML bar on the left provides navigation among the study plan elements. 

Input mechanisms for participants are being worked out.  Initial feedback can be provided to the study leaders in any format (best to use line numbers please):  or

Guide to different sections of the study plan

The management goal framework is developed to set up the study’s key question and geographic and disciplinary scope (Fig. 1).  Different perspectives on the goals and approach are provided by participants in the goals forum. 

Management strategies to achieve the goals will be compared using a scientific framework, where each strategy is applied on a series of replicated experimental units (watersheds).  A watershed strategy (experimental treatment) will include combinations of un-manipulated and manipulated (management treatment) areas across each watershed.  Management strategies and manipulations are not referred to as treatments to avoid confusion.  The chosen watersheds were drawn from type-3 watersheds (basins that drain into the smallest fish bearing streams) that were over 500 acres.  The characteristics and selection process are described in the watershed population section.  

The initial examination of the distribution of state trust lands within the OESF showed that majority of the large watersheds, which are entirely or mostly managed by DNR, are located in the Clearwater and Goodman River landscapes in Jefferson County. This area was selected for the experimental study (Fig. 2).

The study seeks to create sufficiently large differences in responses through the design of individual strategies to increase the chances of achieving statistical significance within the first decade.  For this reason, 15% thresholds were identified in manipulations.  A simple randomized block design is used to address differences between strategies. Four blocks (replicates) were identified each with 4 watersheds, resulting in a total of 16 experimental units.   The final watershed population was based on ecological, proximity, and managerial criteria.  Random selection of strategies within blocks determined the final experimental array (Fig. 3).

Four additional areas were selected for supplemental stand, small tributary studies not requiring watershed responses. These units are not true watersheds as they have surface inlet(s) and are not topographically defined. For this reason, they are not appropriate to asses hydrologically-based responses. 

Multiple management manipulations will be implemented under each of the three active management strategies. They will be drawn from portfolios of silviculture and operations.  All manipulations will be implemented in an operational setting as part of the DNR timber sale program for this district. Along with treatment design, the sustainable harvest level and other factors including staffing and markets will determine how much volume will actually be harvested in the experimental units during the first decade.

Strategies are long-term, hopefully extending well beyond the first decade of commitment by DNR at this point.  Manipulations are not one-time events; they are ongoing to implement strategies and may vary some through time as adjustments are needed to meet strategy objectives. 

The broadest response variables and ideas about strategies and manipulations flow from engagement of local communities, scientists, and professionals, with input from other stakeholders.  This activity is focused on the study area, which is defined as a rural ecosystem in the NW portion of the Olympic Peninsula (Fig. 1), using a portfolio of social science.  Specific response variables are under development in the monitoring section. 

Goal framework and study area

The broad goal framework we have adopted can be called: ecosystem sustainability of peopled landscapes. This framework applies equally well to rural and urban areas (Fig 1).  The goal is to help rural and urban communities sustain themselves and the environment that surrounds them.  These communities and their environment are considered together as a human ecosystem (Burch et. al. 2017). 







Our underlying premise is that communities and their environment are interacting and ultimately interdependent—sustaining one and not the other will not work over the long run.  Aptly described as:

To care for the place, you have to care for the people, and vice versa
– DNR Lands Commissioner, Hilary Franz [June 8, 2017]

To better understand how to achieve this goal, we must study interactions in a defined ecosystem (with a specified boundary) that contains people in a context of surrounding land (for rural areas, mainly forested watersheds).  Our study ecosystem is defined as the NW portion of the Olympic Peninsula (Fig. 2). This ecosystem has already been designated as the hydrologically defined boundary of the Olympic Experimental State Forest.  It contains a variety of small towns, reservations, and lightly populated lowlands as well as small woodlands, industrial forests, publically managed state and federal lands, and parts of the Olympic National Park.  This is an iconic landscape of some of the most productive and beautiful forest lands, streams and rivers, mountains, and ocean beaches in the country.

Applying a specific boundary to this NW Olympics rural ecosystem comes from the tradition in ecosystem science to draw ecosystem boundaries for the purpose of analysis.  The choice of a boundary is known to determine the success of the study. Included in the ecosystem boundary are people living in encompassed rural communities and conceptually also people with major stakes in the area, even though they do not physically reside there.  Including people follows from the goal framework and refocuses on the response of the whole (people and land together).  Homogeneity of social and ecological processes is also important to reduce some of the complexities (hence our use of political and hydrological boundaries). 











This goal and concept can be scaled up to regions that have both rural and urban areas, countries, and the globe.  It cannot be scaled down to the point that human communities or their environment are not included.  Another key concept of ecosystems is that it reflects more than the sum of its parts.  You cannot, for example, sustain a single or small set of species and assume that you are sustaining others species or other values.  The interpretation of individual species or person for that matter has to be placed in the broader context of the system as a whole.  Piecemeal approaches fail in these complex systems.

Core question

This goal framework and study area leads to a specific core question for our study,

Will a higher, sustainable level of rural ecosystem wellbeing2 emerge
from an array of land management strategies implemented
and compared across the OESF landscape?

The “higher level” of sustained wellbeing refers to a yet-to-be-tested assumption that the current level is below what is possible.  This assumption requires including a treatment where higher levels are attempted.  It is also possible that current levels are not sustainable as well, and another treatment is needed to addresses this possibility.

Clear measures of success are needed to form any data-based conclusions about levels of sustainability achieved in this ecosystem.  The most general concept of success we use is ecosystem wellbeing, including both environment and community wellbeing and their interactions (Fig. 1).  Defining and weighing measureable elements of community, environmental, and ecosystem wellbeing is no easy task.  What we do know is that ecosystems of this size are inherently complex and have unique features both in terms of communities and environment—to the point that local knowledge is essential.  This alone suggests a role for locally informed people, scientists, and professionals in defining site-specific wellbeing targets and strategies to achieve them.  Differences in values and uncertainties will not allow for complete agreement on wellbeing targets, but data-based evaluation of a range of targets and sums will greatly improve the debate and, we assume move decisions toward a more sustainable future.  Better understanding how collaborators perceive the goal and wellbeing measures is important to understand how they will interpret outcomes.


Study goals forum


Here we open a neutral forum on study goals from a variety of collaborator and stakeholder perspectives (individually submitted), starting with the organizations providing land and leadership to this study. All stakeholder inputs here and elsewhere will be carefully reviewed by the study team of scientists and professionals, who are responsible for the final content of the plan.    

Washington DNR. As land manager of the Trustlands on which this study will take place, DNR has taken on a variety of goals, and these are enumerated in planning and decision documents including the Habitat Conservation Plan (WADNR 1997) and the OESF Forest Land Plan (WADNR 2016).  The latter specifically establishes an adaptive management process which will be applied to this study.  DNR goals for state trust lands can be summarized as (1) meeting beneficiary revenue mandates; (2) complying with HCP requirements for ESA species and fish, and (3) listening and making changes to accommodate broad stakeholder needs while meeting 1 and 2 above.  The DNR management fee (about 28% of gross revenue) is held constant making efficiency of operations paramount; this in turn can limit learning about new approaches from DNR sources.

University of Washington - ONRC. The ONRC was established with a goal of finding new ways to manage forests that provide conservation and economic benefits at the same time.  A key idea of this study is to develop and compare innovative approaches in ways that can influence future decisions.  The ONRC goals can be summarized as: (1) articulating a new rural ecosystem philosophy to underpin sustainable management of natural resources to benefit both people and place as a whole; (2) working together to develop a model system on the Olympic Peninsula that can identify bottlenecks to sustainability; (3) advancing science-based learning as a third leg of sustainability; and (4) providing a neutral forum to discuss and solve problems.  The learning potential that ONRC provides meshes with the field operations capacity of DNR to help both entities achieve long-term goals.

Rural community stakeholders.  [add input from Forks and Clallam County officials; and/or synopsis of what we heard]

Environmental community stakeholders.  [add input from OFCO, WEC, …; and/or synopsis of what we heard]

Non- profit land manager stakeholders [add input from TNC and HRT; include]

Forest industry stakeholders.  [add input from AFRC, others?]


PNW Station (Anderson)

Olympic National Forest

Olympic National Park


Wellbeing concerns and indicators


The need for site-specific definition of wellbeing drives the development of these targets.  Multiple sources of knowledge play a key role and include research and management- and regulatory-agency scientists who have worked with these or similar forested watersheds, tribal resources specialists, and community members.  This knowledge is fairly diffuse and we begin with easily available sources.  Ecosystem services concepts provide one approach to identifying critical basic services that provide outputs and outcomes that people value more directly.  We start with a list of broad environmental and community targets that we assume are high priority of local communities [subject to addition/change].



Wellbeing concerns


at watershed or larger scale:


CO2 sequestration from the atmosphere

Standing C pools (dNEP)

Soil productivity (indicators and actual growth responses)

Soil C changes (LTEP study and potential NPP model)

Stream “health”

Habitat such as in-stream wood, stream temperature, shade; productivity indicators such as stream nutrients, macroinvertebrates, trophic chains

Late-seral bird habitat (owls, murrelets)

DNR’s definitions in the habitat conservation plan

Early-seral neo-tropical bird habitat

Lidar model? Sonic recordings

Ungulate habitat

Lidar model?

Viable salmonid populations

Fish abundance, distribution and biomass

[add here]



Local jobs

Numbers and distributions by type (especially year-round, family wage)

Local salaries gross/net

Numbers and distributions by type

Revenue to beneficiaries

Timber volume consistent with sustainable harvest levels

Human health

Standard records and trends analyses

 Poverty level

% people eligible for food stamps Standard records and trends analyses

High school students wanting to leave

New survey vehicle?

[add here]



Solar energy capture (photosynthesis driving all foodchains & wood production)

Net primary production (dNPP) and potential NPP model

People-land connectedness

New survey vehicle?

[add here]



 [Please continue list; priority will be assigned later]


Watershed management strategy portfolio

The study proposal (ONRC web ref) developed 4 management strategies in response to the goals and core question.  The brief logic for these strategies is described below; more detailed rationale is in Appendix 2.  The stand-scale practices and landscape designs implementing them are to be fully developed through this study plan.  Near-term implementation is described in Appendix 1.

1.     OESF Forest Land Plan (i.e. DNR’s integrated management strategy). The OESF Land Plan attempts to move away from fixed boundaries of the zoned approach.  For example, old-forest habitat is capped at 40% of the landscape, so as more habitat is achieved through growth, older forests can be cut.  [add more here, including riparian and “sustainable harvest” target]

2.     Zoned Management.  Most public lands (DNR outside of the OESF and the Forest Service) have adopted a zoned or fixed land-use allocation model, where a single or dominant use is assigned parts of the landscape.  These designations are generally unchangeable.  Advantages including ease of oversight have made this the dominant model, but we also know that forests are very dynamic and many species respond over the long run to change more than stability.  Total decadal harvest is expected to be 15% lower than the OESF land plan. 

3.     Accelerated Integration. A variety of new silviculture and operational approaches (at stand and landscape level) are applied in an attempt to increase the level of environmental and community wellbeing with few restrictions on where management can occur (narrow buffers around fish-bearing streams and existing high-quality old growth habitat remain off limits).  Harvesting is expected to be 15% higher than the OESF land plan.

4.     No-Action Control. This strategy was added mainly to examine background changes, especially those resulting from natural disturbances, such as tree mortality from self-thinning, disease, and wind storms.  It also provides more contrast to help detect strategy effects sooner on manipulated watersheds. This strategy is limited to a 10-year period as it does not comply with Trust mandates that require revenue generation from these lands.  No-action, however, is under discussion as a management strategy in Marbled Murrelet blocks by the Board of Natural Resources [see video Here]. Some stakeholders have also argued for carbon parks without harvesting—so, this strategy can be viewed in multiple ways. 

Details of specific silvicultural and operational prescriptions for these strategies are yet to be fully developed.  Initial guidance (Appendix 1) is issued here so that managers will not have to delay ongoing planning which takes years before actual implementation.  This provides a place to begin thinking how strategies can be applied to individual watersheds.

The population of watersheds to be studied

Using DNR GIS data, DNR and ONRC project staff examined about 50 Type-3 watersheds were for possible inclusion in the Jefferson County portion of the OESF using the following size, administrative, and ecological screens:

·      Greater in size than a section (>500 acres) to address landscape questions;

·      Having some older-forest habitat (frequently 1921-blow origin stands);

·      Having some steep and modeled unstable riparian areas;

·      Proportion of young plantations (more recent harvest); and

·      Having a large majority of DNR ownership;

A second set of practicality screens was then considered:

·      Sold and planned harvests where ground data has already been collected;

·      Roading costs to access watersheds further from the currently open system.

Some watershed boundaries were truncated by moving the pour point upstream to avoid non-DNR lands and sold or planned harvests.

With a final set of acceptable T3 watersheds, five groups (blocks) of 4 initially similar watersheds were chosen based on a similarity analysis (Bormann et al. 2008). The main similarity variables we used were the screens above along with proximity.  In the 4 blocks, which contained the 16 hydrologically true watersheds, the 4 management strategies were randomly assigned by roll of the dice by Olympic Region Manager, Mona Griswold (protocol available upon request).

The final set of watersheds (Fig. 3) is distributed throughout the Jefferson County area.  The idea of using watersheds with habitat and steep-slope deferrals as experimental units is to set up microcosms of the entire OESF.  Harvests and other stand-scale activities will occur in the landscape context—that is, their frequency and distribution in the watershed is used to meet OESF management objectives. 

Figure 3. Experimental watersheds as officially designated. Capital letters denotes block (A, B, C, and D); lowercase letters are the treatment (a-accelerated; c-control; p-plan; and z-zoned). Numbers are watershed id. See close-up of one watershed (Fig. 4) and full suite of maps [Here]

Current forest conditions [subject to new info]

The majority of the watersheds were harvested in the 1970s and 1980s, planted, and tended, have developed overwhelmingly into dense conifer stands today.  [add age histogram] Most of these are now 40 to 50 years old with commercial size trees and volumes.  Some eastern, higher elevation stands are not quite to commercial size and need expensive road repairs to access.

Stands were often planted mostly with Douglas-fir, but are usually dominated by hemlock or hemlock/silver fir. Sitka spruce is found sporadically mainly in the river stream valleys and close to the coast.  Redcedar is present is present in small numbers, mainly in the understory of most stands. Red alder was methodically removed from most stands, but can be scattered in patches usually along streams, wetlands and roads. The extent that alder was replaced by conifers in intermittent stream riparian areas in the 1970s harvests is unclear (retrospective study needed).   About X% [where to get definitive data for OESF or JeffCo part?] of stands have not been entered, with the majority of these originating or affected by a severe 1921 wind storm known as 21-blow.  The 21-blow stands typically are dominated by hemlock trees, now nearly 100 years old.  Some residual trees can be older.  Patches of disease mortality and wind disturbance are emerging in these stands in places.  [More on “true” old growth and riparian vegetation?]

Current road conditions [subject to new info]

Roads were largely built in the 1970s to facilitate harvesting.  Most watersheds have ridgetop roads that connect to valley bottom roads down ridge lines.  Spur roads also drop down within the watersheds to landings where trees were cable yarded.  Most current problems are found in road sections connecting valley roads to ridge roads on side slopes.  Problems range from landslides, blownout culverts, to eroded surface rock insufficient to support full log trucks.  Projects are located and scheduled in part to pay for road maintenance as projects unfold.   Road investments are based on strict standards which might be a question (are these working and is more being done than needed).  Other ideas (temporary wooden mat systems for spurs) could be considered.

Figure 4. Example map of experimental watershed—Block B, Control (Bc).  Other watersheds can be viewed on line [Here]


Social science portfolio


The concept of rural ecosystem sustainability placed equal emphasis on community and environmental wellbeing.  This implies that social science needs as much as attention as environmental and other physical sciences, and that the social sciences involved should be participatory as well as observational.  Ecosystem analysis requires local knowledge of both environmental and societal conditions and change mechanisms.   This approach suggests some example high-level social science questions related to the core question that could be applied in this ecosystem:

·      What aspects of forest management are really important to people?

·      Who gets to define community wellbeing?

·      How can DNR management decisions improve community wellbeing? 

·      How do community and environmental wellbeing interact?

·      How important is perception of improving trajectory to investment (e.g., personal, institutional, business)?


Informal interactions leading up to this study plan suggest the following priorities among various community groups and members:

·      Maintaining the land that local people love and depend on;

·      Active management and tourism supporting jobs and community services;

·      Maintaining key manufacturing infrastructure to support markets for wood products;

·      Rediscovery of ethnobotany (tribes);

·      Conservation of wildlife, old-growth and biodiversity as whole

·      Supporting hunting and fishing; and

·      Wilderness values.


Priorities are not the same across and even within groups, but some common ground is possible. Whether common ground can be expanded through study and interaction remains a question.  A wide array of approaches could be applied to approach these questions, for example:


Find venues to talk and listen to diverse sets of stakeholders (1st step study proposal)

Explore in multiple ways what community wellbeing means to people so changes can be monitored

Discover local knowledge and fit it into silviculture and operations (example apply ethnobotany through silviculture)

Work with groups on side projects that connect to the study, if possible

Facilitate written input to the study plan, monitoring, and analyses

Have community participate in all aspects of the study

Involve K-12, and 2- and 4-year college as well as graduate students

Scale up to include linked rural-urban ecosystems

Learn together how to do all of the above and report on it to others


[Study plan authors—Marc Miller and others—will insert more here]


Silviculture portfolio


Management strategies are developed for the watershed as a whole. Silvicultural manipulations, however, are applied at the stand or reach scale on parts of watersheds, except Controls, to achieve strategy objectives. The extent, distribution and type of stand-level manipulations across watersheds affect the watershed cumulative responses. This table describes standard DNR practice and new ideas for stand-scale prescriptions that can be experimentally compared as part of watershed management and as independent stands





Un-aided development is the predominant silvicultural prescription designed to allow stands to develop to a desired condition by themselves.  It may be ideal for long periods in older forests and un-roaded areas, for C sequestration or other purposes, or it may simply result from inadequate funds for stand intervention.


Variable retention harvest is a type of stand-replacement harvest in which

key structural elements of the existing stand (e.g. snags, structurally unique

and other leave trees, down wood) are maintained while a commercial forest stand cohort is re-initiated. Harvests to be followed by later PCT (below).



VRHstd is modified to bring in and extend early seral elements through alder

retention and pre-commercial thinning of hemlock.



Convert to planted red alder in 25 to 30 year rotations following VRHstd mainly in stream-influence areas to increase nutrients and energy into food chains for young salmon and possibly increase revenue.



Commercial thinning, often implemented on the OESF as variable density

thinning, which includes objectives for revenue and habitat complexity. It

may be followed later by wider thinning or VRH depending on specified




Commercial thinning, typically leaving less trees/acre than Thinstd and adding

silviculural treatments for increasing the wind firmness such as edge

feathering and favoring certain tree species and shapes. Predominantly conifer regeneration left to develop into replacement old-forest habitat without further entry in most cases.


Thinwide with alder under-planting to provide early-seral habitat (neotropical

birds, ungulates, fish, and others) followed variably depending on objectives.


Thinning mainly to reduce density quickly to achieve owl habitat status and achieve the 40% landscape threshold.

Thin specialty

Highly limited selective product harvests for revenue and habitat objectives.



Pre-commercial thinning on existing stands less than 20 years old to set up

older stand treatments (above) to variably alter stand density and composition.


Add culturally important, economically viable species in openings to augment production of secondary forest products that may diminish as tree canopy closes (example, beargrass).  This is a special case of VRHes and Thinwide.

Add new

Continuous cover forestry?

[Bill Wells, Drew Rosenbalm, Dan Donato, Greg Ettl, Richard Bigley, and others—will insert more here]


Operations portfolio

Standard DNR contractor and new operations are applied to implement silviculture and other objectives.  Because road building, harvesting, and yarding are expensive, they often strongly influence net revenue from log sales.  Net revenue is much lower in thinnings because of set-up costs relative to harvest per acre compared to VRH.  Under certain conditions (poor roads, tree volume removals, markets) thinnings may not be economically feasible at all.  Recent innovations in small cable systems, cable-assist mechanical harvesters, and temporary road mats, for example, may present opportunities for improving operations economics.  Planting and pre-commercial thinning are also expensive and possible efficiencies might be found.  For example, a recent MS thesis on the nearby long-term ecosystem productivity study found planting alder as well as Douglas-fir reduced hemlock regeneration and presumably thereby reduced costs of hemlock-oriented PCT. 




Logging systems

Roadside cable thinning, cable-assist harvesting, …

Road building


Road management

Road mats, ..





Data on costs and efficiencies of different operations will be used to evaluate potential effects on net sale revenues. 


[Study plan authors—Bill Wells, Woodam Chung, Jennie Cornell and others—will insert operations study proposals here]


Independent stand-scale comparisons


The variety of stand-scale prescriptions, across different experimental units, will allow us to compare stand-scale treatments as well.  Stands with similar prescriptions (regardless of experimental watersheds and strategy treatments) will form a population that can be compared to different populations in stands with different treatments and those established as controls.  A separate statistical design will be applied.  The supplemental small-scale study blocks (Fig. 3, Fig. 5; about 10,000 acres) provide places to study silviculture and other practices independently of the watersheds.  

Figure. 5.  Example non-hydrologic (linear) T3.


[Study plan authors—Greg and others—will insert multiple stand scale studies here]


Secondary questions and measures

The core question is about ecosystem wellbeing, with environmental and community wellbeing elements.  A variety of secondary questions can be asked.  Some will refer to specific wellbeing elements (see wellbeing concerns and indicators).  Others can be asked about specific policies including the sustainable harvest calculation, road standards, harvesting technology, thinning densities, planting and vegetation management for example.  Most of these questions will unfold as the silviculture and operations portfolios are applied to the watershed treatments.  Some of these will grow into separate study plans. 


A few example secondary questions follow.  


[Add examples for a variety of questions]


1.     What are the effects of VRHalder in intermittent-stream tributaries with projected unstable slope areas (Accelerated integration treatment) on ecosystem wellbeing relative to other approaches?


Comparison: identify suitable conifer-dominated intermittent stream tributaries with some modeled unstable areas in Accelerated Integration T3s and similar tributaries in the other treatments (where VRHalder will not be applied).  Other areas might include no action or possibly a Thinstd or Thines.  Similarity to be based on lidar evaluation of slope and existing vegetation (possibly other factors).  Measure averages that apply to the entire stands and tributary (managed or not).   


Near-term measures to include:

·      Net revenue based on actual conifer harvesting and yarding (near stream) and planting costs;

·      Projected future revenue based on net present value accounting for projected competition and density management, alder harvesting, and prices across a range of interest rates; and projected direct and indirect jobs, average annual income, and county and state taxes.

·      Slope movement after the fact based on repeat aerial photo and lidar.

·      Pre and post-disturbance stream temperatures continued as long as a differential is observed, using in-stream sensors.

·      Insect mass (Bernoulli funnel transect) and neotropical bird presence (perhaps through remote microphone tehnology); possibly e-DNA?

·      Larger T3 response (measures at T3 pour point).


Longer term responses to include tree growth, soil movement, PC thinning costs, …


[We only need a small set of secondary questions here, mainly as examples, for the main study plan.]


Statistical models

Watershed-scale responses

Standard ANOVA is planned for comparing watershed-wide responses of ecosystem wellbeing targets across the different experimental strategy treatments (Table 1). No change may or may not be the null hypothesis, depending on intent of the strategy.  Alpha level will likely be 0.05 unless an argument can be made that inherent variability is higher than expected in part due to practicality concerns, where alpha 0.10 will be used.

Table 1.  ANOVA for comparison


Degrees of freedom

of difference between strategies










The experimental design should allow us to make the following contrasts:

·      Integrated Management vs. Control

·      Accelerated Integration vs. Control

·      Zoned Management vs. Control

·      Integrated Management vs. Accelerated Integration

·      Integrated Management vs. Zoned Management

·      Accelerated Integration vs. Zoned Management


Stand-scale and linear responses

Regardless of strategy applied, stand-scale areas and stream and road reaches within a watershed treated with a common silviculture or operations activity create a larger population of stands, tributaries, and road segments where different activities or lack of activity can be compared (see portfolio pages).

Variable-scale responses

Scale-based response thresholds may be examined where, for example, response below a harvested stand can be tracked downstream until it disappears.

Monitoring philosophy and minimums

Far too many measures than could be afforded that are needed to meet typical comprehensive assessments of ecosystem services, or any likely set to please all researchers or other study participants. Choosing what to monitor will have a lot of influence on success of this study. Setting priorities will be a group effort. Several criteria are proposed to help make these choices:


·      Perceived relationship to future decisions under the control of DNR;

·      Connection to specific wellbeing elements;

·      Distribution across wellbeing elements (environmental, community, ecosystem); and

·      Cost (zero-sum game).

Response variables will be chosen primarily to inform potential future decisions. For example, response variables could be net revenue return of wide thinnings, the time to achieve habitat structures in owl thinnings, and the cost of maintaining roads in each experimental unit. Monitoring metrics that cannot be clearly tied to management decisions or that are too expensive will be avoided. That is, monitoring resources will be focused on how management is directly affecting the bases for deferrals (for example, evidence of accelerated development of spotted owl, murrelet, or riparian habitat) but also evidence of changes in other measures of environmental and community wellbeing (for example, distribution of seral stages, jobs, and accounting of relative management costs and beneficiary revenues). Future decisions can then be based on the entire range of treatment effects.

We propose to use detailed activity accounting and remote sensing (repeat LiDAR and aerial photos) as the backbone of the monitoring system. LiDAR can be linked to limited ground measures to evaluate outcomes on a wall-to-wall basis. Field measurements will be used when remotes sensing data collection is deemed inadequate.

Retrospective analyses will be added, if possible, to extract knowledge from past actions. Modeling will be added to guide the study design and implementation. The environmental and economic outcomes of implementing the OESF Forest Land Plan treatment have already been modeled as part of the EIS analyses. Modeling the outcomes of the other two strategies can give us a treatment schedule for each of the experimental units based on an optimal model run. The empirical data collected post-treatment will be used to assess the model projections.

Funding for monitoring has yet to be fully acquired.  It may come from beneficiaries of DNR management, grants, and other sources.  A goal is to create an endowment that can assure funding over time commensurate with the study design.


[Lots of work to do here]


Work completed (under development)

Status Report – September 18, 2017


Watersheds were designated by randomly assigning treatments on September 14, 2017 (Table 1, Fig. 3).  Experimental units are referred to as (BLOCK treatment).  A “9-“ in front of the 3-digit watershed number (known as WSID) indicates that boundaries were redrawn to exclude planned activities below a new pour point.  Boundaries from the original DNR T3 polygons are also being redrawn more accurately using lidar data. 


Table 1.  Official T3 watershed designations (randomly assigned within blocks)


Experimental Treatments







9-778 (Ac)

9-789 (Ap)

9-786 (Az)

790 (Aa)


565 (Bc)

9-702 (Bp)

587 (Bz)

748 (Ba)


654 (Cc)

784 (Cp)

9-729 (Cz)

9-660 (Ca)


9-744 (Dc)

787 (Dp)

758 (Dz)

780 (Da)

§ T3 720 stands as an alternate watershed if a problem arises with another C block unit


A group of 4 additional areas consisting of smaller watersheds mostly along larger rivers systems was also identified as part of the study.  These were designated as Block L (Fig. 5).  These areas were set up to facilitate stand and small-watershed studies not possible in the main part of the experiment.  No random designation was needed at this time.  We envision drawing numerous, smaller experimental units from these areas and randomly assigning treatments at a later date.


A series of blocking schemes were examined before these T3 sets per blocks were agreed to (balancing science and management needs).  A major effort was given to finding sets where each T3 within a block would be acceptable as a Control.  Active and planned timber sales made this difficult.  The adopted scheme blocked primarily by proximity.  Block A units lie on the south side of the Clearwater river; Block B are north-central within the Jefferson County DNR lands; Block C units are on the eastern portion with drainages flowing to the west; and Block D are more coastal.  Average size was similar across blocks while percent of old forest ranged from 12 to 28% across blocks (Fig. 1-1).  Looking at experimental units (T3s) selected by experimental treatments, Zoned and Accelerated treatments tended to be in larger watersheds but old forest% did not differ much (Fig. 1-2). 


A few issues remain to be resolved.  A minor boundary adjustment may be needed in Ac to avoid a small portion of the timber sale “Bad manors.”  Work is needed to refine silviculture activities in the planned sales to assure they are compatible with the Zoned treatments Az and Cz. 


Timelines (under development)


Benda, L.E.; Miller, D.J.; Dunne, T.; Reeves, G.H.; Agee, J.K. 1998. Dynamic landscape systems. In: Naiman, R.J.; Bilby, R.E., eds. River ecology and management: lessons from the Pacific coastal ecoregion. New York: Springer: 261–288.

Bigley, R.and F. Deisenhofer. 2006. Implementation Procedures for the Habitat Conservation Plan Riparian Forest Restoration Strategy. Department of Natural Resources, Olympia, Washington

Bormann, B.T. , B. K. Williams, T. Minkova. 2017. Learning to learn: the best available science of adaptive management. In: Van Horne, B. and D. H. Olson, Editors. People, Forests and Change: Lessons from the Pacific Northwest.  Island Press.

Bormann, B.T. J.A. Laurence, K. Shimamoto, J. Thrailkill, J. Lehmkuhl, G. Reeves, A. Markus,D.W. Peterson, and E. Forsman. 2008. A management study template for learning about postwildfire management. Gen. Tech. Rep. PNW-GTR-777. Portland OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. 27 p.

Bormann, BT, and MG Kramer. 1998.  Can ecosystem process studies contribute to new management strategies in coastal Pacific Northwest and Alaska?  Northwest Science 72:77-83.

Bormann, B.T., and A.R. Kiester. 2004. Options forestry: acting on uncertainty. Journal of Forestry 102: 22-27.

Bormann, B.T., R.W. Haynes, and J.R. Martin. 2007. Adaptive anagement of forest ecosystems: some rubber hits the road? BioScience 57(2):187-192.  Island Press.

Bunnell, F. L., and G. B. Dunsworth (eds.). 2009. Forestry and Biodiversity. Learning How to Sustain Biodiversity in Managed Forests. University of British Columbia Press.

Carey, A. B., C. Elliot, B. R. Lippke, J. Sessions, C. J. Chambers, C. D. Oliver, J. F. Franklin, and M. G. Raphael. 1996. Washington Forest Landscape Management Project – A Pragmatic, Ecological Approach to Small-landscape Management. USDA Forest Service, Washington State Department of Fish and Wildlife, and Washington State Department of Natural Resources, 110 p.

Commission on Old Growth Alternatives for Washington’s Forest Trust Lands. 1989. Final Report submitted to Bryan Boyle, Commissioner of Public Lands. Washington State Department of Natural Resources, Olympia, Washington.  DNR – refer to Washington Department of Natural Resources

Donato, D.C., J.L. Campbell, and J.F. Franklin. 2012. Multiple successional pathways and precocity in forest development: can some forests be born complex? Journal of Vegetation Science 23: 576-584.

 [FEMAT] Forest Ecosystem Management Assessment Team. 1993. Forest Ecosystem Management: An Ecological, Economic, and Social Assessment. Portland (OR): US Department of Agriculture, Forest Service, US Department of Commerce, National Oceanic and Atmospheric Administration, US Department of the Interior, Bureau of Land Management, US Fish and Wildlife Service, National Park Service, Environmental Protection Agency.

Franklin, J.F. 1993. Preserving biodiversity: species, ecosystems, or landscapes? Ecological applications 3: 202-205.

Franklin, J.F., et al. 2002. Disturbances and structural development of natural forest ecosystems with silvicultural implications, using Douglas-fir forests as an example. Forest Ecology and Management 155: 399-423.

Franklin, J.F.; Johnson, K.N. 2012. A restoration framework for federal forests in the Pacific Northwest. Journal of Forestry. 110(8): 429–439.

Harrington, C.A., Roberts, S.D. and Brodie, L.C., 2005. Tree and understory responses to variable-density thinning in western Washington. In: Peterson, C. E.; Maguire, D. A., eds. 2005. Balancing ecosystem values: innovative experiments for sustainable forestry: Proceedings of a conference. Gen. Tech. Rep. PNW-GTR-635. Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. 389 p.

Holling, C.S. (ed.). 1978. Adaptive Environmental Assessment and Management. Chichester UK: Wiley.

Lee, K.N. 1993. Compass and Gyroscope: Integrating Science and Politics for the Environment.  Washington, DC: Island Press.

Murphy, D.D., and B.R. Noon. 1992. Integrating scientific methods with habitat conservation planning: reserve design for northern spotted owls. Ecological Applications: 4-17

Olson, D. H., and Van Horne, B. Editors. 2017. People, forest, and change: Lessons from the Pacific Northwest. Island Press.

Pickett, S.T.A and P.S. White. 1985.  The ecology of natural disturbance and patch dynamics.  Academic Press, San Diego, CA

Reeves, Gordon H.; Pickard, Brian R.; Johnson, K. Norman. 2016. An  initial evaluation of potential options for managing riparian reserves of the Aquatic Conservation Strategy of the Northwest Forest Plan. Gen. Tech. Rep. PNW-GTR-937. Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. 97 p.

Shaffer, M.L. 1981Minimum population sizes for species conservation. BioScience 31: 131-134.

Van Mantgem, P.J., et al. 2009. Widespread increase of tree mortality rates in the western United States. Science 323: 521-524.

Washington Department of Natural Resources. 1997. Final Habitat Conservation Plan. Washington State Department of Natural Resources, Olympia, Washington. [;]

Washington Department of Natural Resources. 2004. Final Environmental Impact Statement on Alternatives for Sustainable Forest Management of State Lands in Western Washington and for Determining the Sustainable Harvest Level. Washington State Department of Natural Resources, Olympia, Washington.

Washington Department of Natural Resources. 2006. Policy for Sustainable Forests. Washington State Department of Natural Resources, Olympia, Washington.

Washington Department of Natural Resources. 2016a. Olympic Experimental State Forest HCP Planning Unit - Final Environmental Impact Analysis. Washington State Department of Natural Resources, Olympia, WA

Washington State Department of Natural Resources. 2016b. Olympic Experimental State Forest HCP Planning Unit Forest Land Plan. Washington State Department of Natural Resources, Olympia, WA. []

Walters, C.J. 1986. Adaptive Management of Renewable Resources. Caldwell NJ: Blackburn Press.

Wilcove, D.S., C.H. McLellan, and A.P. Dobson. 1986. Habitat fragmentation in the temperate zone. Conservation biology 6: 237-256.

Wondzell, S.M.; Hemstrom, M.A.; Bisson, P.A. 2007. Simulating riparian vegetation and aquatic habitat dynamics in response to natural and anthropogenic disturbance regimes in the Upper Grande Ronde River, Oregon, USA. Landscape and Urban Planning. 80: 249–267. doi:10.1016/j.landurbplan.2006.10.012.



Glossary (add as needed)
(MW refers to definitions from [])


Community.  MW: 1: a unified body of individuals; 2: society at large. By itself this conflicts with ecologists usage as animal community, so requires an adjective for them if out of context. Conclusion: critical word to show this is about groups not individuals. 

Community wellbeing.  Given the definitions of included words, this phrase should be clear as the condition (health, prosperity, happiness) of a human community (place, group, or society at large).  We use this to express the breadth we seek (expanding from simple economic vision of the past).  It is sufficiently vague so that people can agree but then must define, for themselves. Quality of life can be used as a surrogate, as long as it is not referring to an individual.

Ecology.  MW: a branch of science concerned with the interrelationship of organisms and their environments. Alternative definition is: biota-environment interactions. Vernacular uses abound such as a social movement about environmental protection.  Conclusion: avoid because it means non-human world to some people.

Ecosystem. MW: the complex of a community of organisms and its environment functioning as an ecological unit.  Alternative definition is: a biological community of interacting organisms and their physical environment. Vernacular: complex network or interconnected system.  Business: interaction of your business with its surrounding entities. By itself this conflicts with some ecologists use.  Conclusion: required since it is at the core of our approach but must be combined with other words.

Ecosystem sustainability.  To continue the desired condition of a defined ecosystem for a long period. By itself this conflicts with some ecologists use as non-human ecosystem.  See Rural/urban ecosystem sustainability. Conclusion: combines succinctly 2 key concepts, but needs modifiers to connect to nontechnical audience.

Environment.  MW generally: circumstances, objects, or conditions by which one is surrounded.  MW science focus: the complex of physical, chemical, and biotic factors (such as climate, soil, and living things) that act upon an organism or an environmental community and ultimately determine its form and survival.

Environment wellbeing.  Given the definitions of included words, this phrase should be clear as the condition of the environment surrounding human communities.  Conditions include the biotic and abiotic surroundings, including individual species, biotic communities of species, and biogeochemical processes such as photosynthesis, C sequestration, nutrient cycling, erosion, weathering, and hydrology, for example.  Disturbance frequency and intensity are also included.


Human ecosystem.  [add ref to Burch et. al. 2017]

Nature.  Relating to our focus, MW: 6: the external world in its entirety. Conclusion: many people think this means the non-human world (i.e., natural); while others such as Native Americans find the distinction between people and nature invalid—because of this conflict, we avoid its use.

People and place.  The DNR Commissioner of Public Lands used this phrase in front of a large audience of WA citizens recently and it seemed to work well as an alternative way of explaining the rural ecosystem sustainability model.  .

Rural/urban.  The US Census has a quantitative definition [web ref]. Clearly rural and urban can be combined to from regional or larger ecosystems where rural-urban interactions could be the focus of study.

Sustainability.  MW: 2a: of, relating to, or being a method of harvesting or using a resource so that the resource is not depleted or permanently damaged. Perhaps better stated as, maintaining or improving something over a lengthy but defined period.  The something is almost anything and this distinction is critical (can be human or other animal communities, ecosystem process, timber supply, mill profits, …). Conclusion: this word is required since it is at the core of our approach but it must be combined with other words.  This comes in part from sustainable development concepts of World Bank and others.

Wellbeing.  MW: the state of being happy, healthy, or prosperous.  We argue it’s more than a yes/no state, with multiple levels of wellbeing.  Many probably begin with a human health perspective, but health has been widely applied to natural resources.  Wellbeing works better than health as it is broader, clearly including economics.

Win-win.  This phrase has worked well to help people see value in focusing on both community and environmental wellbeing.


Appendix 1.  Interim guidance to DNR land managers

This experiment is being imposed on top of ongoing management already underway in some cases.  Here we offer guidance to managers for near-term decisions, before the study plan is completed.  This also offers a chance to think about how we can apply existing and new prescriptions and operations to help shape the next interventions.


1.     OESF Forest Land Plan (Plan)


Watersheds so designated are: WSIDs 9789 (Ap), 9702 (Bp), 784 (Cp), and 787 (Dp).  Sold harvest units will go forward as designed (no option to alter extent, location, or prescription).  In-process sales will have the extent altered so as not to exceed the total volume/watershed based on an estimate of the new sustainable harvest calculation for the OESF as a whole (per acre adjusted).  Future harvests planned in the first decade will move the watershed toward this estimate.  We will consider whether to make adjustments based on specific characteristics (deferral and operational constraints) at a later time. 


Across the watershed, standard DNR silviculture (VRHstd, Thinstd, and PCT) will be used in accordance with the forest land plan.  Temporary deferrals for older forest blocks will have NoAction during the first decade, along with other areas not entered during the first decade.  Entry into modelled unstable riparian areas is allowed under the forest land plan and would continue standard practice of conifer-oriented regeneration.


2.     Zoned Management (Zoned)


Watersheds designated as Zoned are: WSIDs 9786 (Az), 720 (Bz), 9729 (Cz), and 758 (Dz).  Sold harvest units will go forward as designed (no option to alter extent, location, or prescription).  In-process sales will have the extent altered so as not to exceed the total volume/watershed minus 15%, based on an estimate of the new sustainable harvest calculation for the OESF as a whole (per acre adjusted).  Future harvests planned in the first decade will move the watershed toward this estimate.  We will consider whether to make adjustments based on specific characteristics (deferral and operational constraints) at a later time. 


A permanent late-successional reserve (LSR) will be designed around the existing old forest, regardless of current habitat status to increase it to 40% of the watershed.  Old forests now occupy about 20% of the watersheds on average.  Multiple decisions are needed, for example whether to block up existing chunks, whether to have stream to ridge blocks, and whether to specifically include/exclude operationally accessible areas.  Riparian buffers, as used by DNR outside of the OESF, will also become permanent reserves.  [we need to decide if they too should be considered part of the LSR, thereby reducing acres above]. 


In Zoned watersheds outside of reserves, prescriptions should favor VRHstd and Thinwide.  Favored prescriptions for younger stands inside late-successional reserves are Thinowl or Thinwide or Thinstd followed 5 or more years later by Thinwide in areas more prone to wind storms.  Thinning is used to speed development of late-seral habitat and as a potential replacement old forest in the event of losing old-forest habitat in the reserved area.  The volume target for Zoned is to harvest at least 15% below that of the Plan target.


3.     Accelerated Integration (Accelerated)


Watersheds designated as Accelerated are: WSIDs 790 (Aa), 748 (Ba), 9660 (Ca), and 780 (Da).  Entire watersheds are open for some sort of active management except for unstable slopes as determined by onsite inspection and [VRH in some sort of fish-bearing stream buffer?].  New silviculture prescriptions will be applied that respond to new ways to provide community and environmental benefit not currently used in standard DNR practice.  For example, some stable riparian areas may have VRHalder or VRHes to increase hardwood shrubs and trees to increase insect and detrital inputs to fuel aquatic food chains.  Upland areas may have ethnoforestry or elk habitat prescriptions, variants of VRHes and standard PCT.  Old forest, yet to achieve late-seral habitat conditions will be open to Thinspecialty to speed development and pay for it through selective thin-from-below and species targeting.  More approaches are expected as the study plan unfolds.  The volume target for Accelerated is at least 15% above the volume targets for Planned.   Note that harvests exceeding the sustainable harvest calculation will be offset by other treatments at a whole-study or OESF basis.


4.     No-Action Control (Control)


Watersheds designated as Controls are: WSIDs 9778 (Ac), 565 (Bc), 654 (Cc), and 9744 (Dc).  These will deploy NoAction throughout for the first decade. 




Appendix 2: Strategy rationales

A debate has been underway in the science and land management communities since about 1990 about the best basis for managing forest lands. Three main approaches have dominated:

·      A tree-farm model (practiced on most private lands) that focuses on renewable wood production with mitigation for habitat and other issues, often just enough to satisfy limited state and federal regulations.  Decisions are based on owner objectives, which can vary substantially among organizations, but typically are to maximize revenue.

·      On most public lands, the current dominant approach uses a conservation_biology basis for management (also known as a zoned or land-sparing approach).  This approach calls for dividing the land into large land_use designations - blocks managed for individual purposes, such as late_successional habitat in reserves or timber production in non-reserved lands (matrix).  Conservation biologists have emphasized habitat reserves as a means to protect endangered species (Schafer 1981, Wilcove et al. 1986, Franklin 1993).  Lawsuits based on the endangered species act provide legal precedent by mandating habitat reserves to support recovery of listed threatened and endangered species (e.g., the Northern spotted owl; Murphy and Noon 1992).  An example is managing the national forests under the Northwest Forest Plan (FEMAT 1993).  One of the main criticisms of the zoned approach is that it ignores temporal dynamics created by natural disturbances—to which most species would seem to be adapted.  Future disturbance can be expected and even exacerbated through climate change, affecting habitat attributes through mechanisms such as increasing tree mortality rates (Van Mantgem et al. 2009), thus requiring recruitment of new suitable habitat.  Late_seral conditions naturally develop over several centuries (Franklin et al. 2002), and key drivers of late successional habitat are not yet firmly established (Donato et al. 2012). Therefore, we cannot be certain that the zoned land management approach is the best way to assure viable populations of listed species.

·      A third approach has long been under discussion, but has not been widely applied.  It is based on disturbance ecology as an alternative basis for management by recognizing a natural array of successional forest stages that shift in space over time through disturbances in a shifting mosaic pattern (Pickett and White 1985). It is also called un-zoned or land-sharing management. A similar paradigm has emerged for streams, viewing them as dynamic in space and time, exhibiting a range of potential conditions, as do the terrestrial systems in which they are imbedded (Reeves et al. 2016 and the references therein). The temporal variability of individual streams and watersheds depend on landscape context (e.g., topography, geology) and past natural and management disturbances (Benda 1998, Reeves 1995, Wondzell 2007).  Management approach for riparian areas under this paradigm is “context-dependent” and most often is expressed as varying the width of the riparian buffers. This new approach for managing uplands and riparian areas seems to have the potential to better integrate timber production and habitat conservation (Reeves et al. 2016) and better consideration of ecological processes that underpin long-term ecosystem productivity (Bormann and Kramer 1998).

When the DNR created the OESF, it decided to implement and evaluate the third approach, naming it “integrated management.” This term was first introduced in the HCP (DNR 1997) and recently refined in the OESF draft Forest Land Plan (DNR 2016b) to mean an experimental management approach based on the principle that a forested landscape can be managed with different level of intensity through time and space to provide both revenue production and ecological values.   Because integrated management was new and has not been tried in the Pacific Northwest before, the DNR recognized early on the need for research, monitoring, and adaptive management to evaluate and continuously improve the approach, and built this into the HCP to assure the HCP goals are met (DNR 1997). The learning effort proposed here supports this objective.



Appendix 3. Review proposed budget and discussion on funding options (under development)



Appendix 4: Estimated workloads and DNR staff needs

(under development)                                 

[1]T3 refers to DNR’s type-3 watershed designation.  These watersheds were drawn to be just large enough to include a fish-bearing stream segment above the pour point.

[2]See Glossary for definitions