EXECUTIVE SUMMARY

Department of the Interior (DOI) bureaus have invested heavily (for example, the U.S. Bureau of Land Management (BLM) spent more than $60 million in fiscal year 2007) in seeding vegetation for emergency stabilization and burned area rehabilitation of non-forested arid lands over the past 10 years. The primary objectives of these seedings commonly are to (1) reduce the post-fire dominance of non-native annual grasses, such as cheatgrass (Bromus tectorum) and red brome (Bromus rubens); (2) minimize the probability of recurrent fire; and (3) ultimately produce desirable vegetation characteristics (for example, ability to recover following disturbance [resilience], resistance to invasive species, and a capacity to support a diverse flora and fauna). Although these projects historically have been monitored to varying extents, land managers currently lack scientific evidence to verify whether seeding arid and semiarid lands achieves desired objectives. Given the amount of resources dedicated to post-fire seeding projects, a synthesis of information determining the factors that result in successful treatments is critically needed.

Although results of recently established experiments and monitoring projects eventually will provide useful insights for the future direction of emergency stabilization and burned area rehabilitation programs, a chronosequence approach evaluating emergency stabilization and burned area rehabilitation treatments (both referenced hereafter as ESR treatments) over the past 30 years could provide a comprehensive assessment of treatment success across a range of regional environmental gradients. By randomly selecting a statistically robust sample from the population of historic ESR treatments in the Intermountain West, this chronosequence approach would have inference for most ecological sites in this region.

The goal of this feasibility study was to compile and examine historic ESR records from BLM field offices across the Intermountain West to determine whether sufficient documentation existed for a future field-based chronosequence project. We collected ESR records and data at nine BLM field offices in four States (Oregon, Idaho, Nevada, and Utah) and examined the utility of these data for the development of a chronosequence study of post-fire seeding treatments from multiple sites and different ages (since seeding) throughout the Intermountain West. We collected records from 730 post-fire seeding projects with 1,238 individual seeding treatments. Records from each project ranged from minimal reporting of the project’s occurrence to detailed documentation of planning, implementation, and monitoring. Of these 1,238 projects, we identified 468 (38 percent) that could potentially be used to implement a field-based chronosequence study. There were 206 ground-seeding treatments and 262 aerial-seeding treatments within this initial population, not including hand plantings. We also located a considerable number of additional records from other potential field offices that would be available for the chronosequence study but have yet to be compiled for this feasibility report.

There are a number of potential challenges involved in going forward with a field-based chronosequence study derived from data collected at these nine BLM offices. One challenge is that not all seed mixtures in ESR project files have on-the-ground confirmation about what was sown or rates of application. Most projects, particularly records before 2000, just list the planned or purchased seed mixtures. Although this could potentially bias assessments of factors influencing establishment rates of individual species for treatments conducted before 2000, a chronosequence study would not be intended to assess success solely at the species-level. Treatment success would be evaluated based on the establishment of healthy vegetation communities, such as the abundance and density of perennial species, regardless of their lifeforms (grasses, forbs, and shrubs) or their origin (native or introduced), relative to invasive annual grasses. A secondary challenge is that most seeding projects conducted before 1999 on either BLM or U.S. Forest Service lands had little monitoring information available compared to more recent projects. Although a chronosequence study would benefit from comparing current vegetation metrics to those collected immediately following the treatments, this may not always be possible. A preference would be placed on those projects with post-treatment monitoring documentation, but this is not absolutely necessary for success of the chronosequence approach. Finally, post-fire management of fire rehabilitation treatments can have significant effects on the condition and persistence of seeded vegetation. Information on post-fire management of treatment areas after the first 2 years largely is lacking in our present dataset. These data would need to be collected for study sites selected for a field-based chronosequence study. Given these caveats, there is still a sufficient population (at least 468) of well-documented post-fire seeding treatments from which to sample and conduct a robust chronosequence analysis of treatment success for the Intermountain West.

In a potential chronosequence study, the population of projects would be stratified into meaningful categories for distinguishing potential thresholds for seeding. The random-stratified sample of treatments that would be reassessed would be drawn from this population. These categories would include precipitation zones (low, <8 in.; medium, 8 to <12 in.; and high >12 in.), major land resource areas (similar to ecoregions, but related hierarchically to soil mapping units and ecological sites; for example, Snake River plains, Malheur high plateau), and timing of seeding (fall, winter, or spring). Once projects to revisit are identified, field sampling would be completed within previously seeded and unseeded areas with the same soils and climate (ecological site) to determine current vegetation and soil stability conditions. For sites with documented ecological site descriptions (as defined by the U.S. Department of Agriculture Natural Resource Conservation Service, the current interagency standard), these site descriptions could provide suggested relative dominance (cover or production) by species for life forms. These descriptions also would be used as a potential standard for comparisons in treated areas. Each site would be characterized according to topography, soils and pre- and post-treatment precipitation. Potential response variables measured at each study site would examine fuel loads and composition (relative cover or production) of all plant species grouped into meaningful categories such as shrub, grass, forb, native or introduced, or invasives. Abiotic response variables, such as surface cover of bare ground and litter, also would be collected. Data on natural variation (for example, post-treatment weather) and land-management activities (for example, livestock and wild horse herd management data) would be used as covariates in analyses. A study design that incorporates gradient analyses via multivariate statistics most likely would be the best approach for the chronosequence investigation.