Scotch Broom

Many lessons were learned from our five-year study at JBLM, ranging from the mundane to the profound. Here we share what we consider to be our most practical messages based on our data and experience studying broom control:

• When estimating broom cover, run a random transect and use line-intercept method. This is the most flexible and repeatable method across a broad range of broom stand ages and densities.

• When estimating seed germination in the field, use narrow belt transects. Seedlings are very patchy at the scale of meters and vary greatly between sites, but are surprisingly consistent at the scale of tens to hundreds of meters across a site.

• When broom was cut with machinery, the height of the stumps did not influence the probability of resprouting.

• After initial broom removal, go into the sites the following summer to assess both the density of germinating seedlings and the frequency of resprouting stumps. These two pieces of information will predict how fast broom will cover the site (germination) and how fast new seeds will be produced (=1-2 years faster when plants come from resprouts). Manual cutting later may also be more effective when the plants are resprouts rather than new plants.

• Targeting the earliest life stages (seeds and seedlings) might appear to be the most efficient way to control broom populations mechanically, but it was not an effective method at JBLM. Multiple years of flushing the seedbank with soil scarification did not reduce broom cover three years later. In addition, the cost of the treatment in terms of labor and broken blades was excessive.

• Sites with high amounts of germination (e.g. Tanktable and Nisqually in our study) can jump to 50+ percent cover of broom within two years. These thick carpets of broom seedlings seem to compete heavily with other vegetation. Broom cover in sites with mostly resprouting stumps can catch up by year 3.

• Herbicide (Garlon 4) was the only treatment that showed any effect on small (1-year-old) seedlings, and the effect was measurable three years later even where spraying was done through thick grass. This technique appeared to be effective in part because it does not require the operator to see the broom seedlings. The use of dye ensures complete coverage.

• The kill rate with 2.5-year-old-plants was more consistently high for chemical than mechanical control. The best results for brush-cutting were as good as herbicide, but most sites showed poor results.

• It is not possible to predict death from herbicide by looking at the plants 6 weeks after spray. We found that % green tissue at 6 weeks did not predict whether the plants were dead 12 months later.

• Taller plants were NOT less susceptible to herbicide (within a year and within a site, on 2.5 or 3.5-year old plants).

• In this study, season was NOT a critical factor in herbicide effectiveness. Spraying in March, May, and September gave similar results. This is an important finding because managers are free to determine when plots are sprayed based on logistical priorities.

• The kill rate for herbicide was as high for 3.5-year-old plants as for 2.5-year plants in some sites, but had dropped from 90% to about 70% in two of the sites.

• The use of Garlon 4 on broom did not seem to harm Douglas fir trees in several different experiments.

• Garlon 4 treatment gave long-lasting results. While control plots at most sites were at 100-150% broom cover by 2012, plots sprayed in 2009 were still at 10-30% cover. In January 2014, herbicide plots were still easily distinguished (see photo below).

• Trees planted near forest edges survive better, even in sites where plantations have failed multiple times. Among several possible explanations, mycorrhizal connections with established trees may be involved.

• Harvest and reforestation strategies should maximize the benefits of forest edges. This has implications for the geometry and optimal size of timber harvests.