Comparison of Non-herbicide Methods for Control of Invasive Bush Honeysuckle as Preparation for a Prescribed Burn

Results from 2024

The goats did an outstanding job in spring 2024 of opening up the woods by defoliating everything to about 6 ft in height. This made subsequent mechanical clearing much easier, because machine operators could see where they were going and manual clearing was less encumbered by brushy undergrowth.

Goats not only ate the honeysuckle leaves, they also ate some of the smaller branches, and they climbed on the trunks and cracked some of them off. Video 1 shows goats working together to stand on honeysuckle and bend it over to eat the leaves.

Video 2 shows a BWWI manager calling goats to a new paddock that we had just completed fencing. Upon reaching the new paddock, the goats immediately started eating honeysuckle; they go right to work once allowed into a new paddock.

Graphs of the main metrics for honeysuckle (number of stems, percent cover, and average height) in the test plots and control plot are shown in Figures 11 through 13. The data are rather noisy, as it is difficult to accurately estimate the metrics. Furthermore, these metrics are not independent variables and do not consistently correlate. Nonetheless, the metrics provide insight into the effectiveness of the treatments in each unit.

As shown in Figure 11, the number of honeysuckle stems did not change much across the four count dates, except in the case of test plot A1a. The last count in the series on September 30, 2024 shows a large drop in the number of stems in test plot A1a due to mechanical clearing. The other three test plots had not yet been mechanically cleared by September 30, 2024.

Goat grazing generally had a large impact on honeysuckle percent cover, that is, percent of the ground shaded by honeysuckle, as shown in Figure 12. Grazing occurred in May and June 2024, and the drop in percent cover from May 8 to July 2, 2024 count dates is significant. The goats thoroughly defoliated the honeysuckle they could reach. Only test plot A1b did not show a significant drop in percent cover, due to large honeysuckle above goat graze height in that area. An additional drop in percent cover is exhibited for test plot A2b in September 2024. Test plot A2b was moderately grazed in the spring and heavily grazed in the fall, leading to an additional decrease in honeysuckle percent cover by September 30, 2024. The untreated control plot A5 shows little change in honeysuckle cover over the entire period.

Figure 13 reveals that spring grazing slightly reduced honeysuckle average height compared with the untreated control. Honeysuckle average height decreased 40% on average in the test plots from May 8 to July 2, 2024 while average height decreased 30% in the untreated control plot A5 during that same time period. Average height in the untreated control plot A5 continued to decrease slightly throughout the year, suggesting a time-of-year effect. Honeysuckle average height rebounded by September 30, 2024 in test plots A1a and A2a. Test plot A1a was mechanically cleared in September 2024 but a couple rather tall honeysuckle remained after clearing, resulting in an increase in average height though far fewer stems (see Figure 11). Unit A2a was not grazed by design in the fall, but neither was the untreated control for which height did not rebound. This suggests that the rebound in honeysuckle average height for test plot A2a may be within the noise in the data.

Figures 11 through 13 show little difference from July 2 to September 5, 2024 for any of the three metrics. There was not much honeysuckle growth over the summer. Weather over the summer was not atypical for the area (CIProud.com, ‘Looking back at meteorological summer 2024’), so lack of growth cannot be blamed on unusual weather. Photos in Figure 14 of the mulched unit A1/2e show thick growth of mostly weeds over the summer, as unit A1/2e was in the sun. In contrast, the other units had very little regrowth over the summer presumably due to heavy shade. Although percent honeysuckle cover was generally low in the woods after spring grazing, the ground was still heavily shaded from midstory and canopy trees.

It was very difficult to walk through the woods before goat grazing due to heavy underbrush, but after spring 2024 grazing, walking (and seeing) through the woods was remarkably easier. Photos taken during the stem counts show the difference pre- and post-grazing in the study areas (Figures 15 through 18). In contrast, honeysuckle in the untreated control area A5 remained thick (Figure 19).

It was hoped that goats could eat honeysuckle berries during fall grazing, because after passing through the goats’ digestive tracts, honeysuckle seeds may no longer be viable. However, we found that the berries were mostly higher than the goats could reach. Generally speaking, tall honeysuckle bushes produce berries, whereas short, shaded honeysuckle bushes do not. Short honeysuckle bushes defoliated by goats in the spring definitely did not produce berries and were even struggling to produce leaves.

Stocking rate varied from about 20 to 100 goats/ac, but this did not appear to impact effectiveness. This result agrees with a study from Purdue University (Rathfon, et al., 2021), in which stocking rate was a controlled variable. Investigators reported that lower stocking rates extended grazing duration, but did not affect grazing effectiveness.

Results from 2025

Graphs of the main metrics for honeysuckle (number of stems, percent cover, and average height) in the test plots and control plot are shown in Figures 20 through 22. Note that all honeysuckle were counted in 2025 whereas only honeysuckle stems larger than 0.3-inch diameter at 18-inch height were counted in 2024, so this must be taken into account when comparing 2024 and 2025 results. For example, the untreated control plot A5 had about 30 more honeysuckle stems in 2025 compared to 2024, primarily due to the change in counting method.

When mechanically clearing 20 acres, it’s easy to miss patches and unfortunately test plot A2b did not get mechanically cleared, but the rest of unit A2b was mechanically cleared. Therefore, the metrics collected for test plot A2b do not accurately reflect unit A2b overall. Note that goats don’t miss patches; they eat nearly all vegetation in reach.

As shown in Figure 20, the number of honeysuckle stems for all test plots was significantly lower than the untreated control plot A5, even for test plot A2b, which hadn’t been mechanically cleared.

Honeysuckle percent cover was about 5% for test plots A1a, A1b, and A2a both before and after grazing, as shown in Figure 21. In contrast, test plot A2b, which hadn’t been mechanically cleared, showed 65% honeysuckle cover in May 2025, but goat grazing over the summer reduced honeysuckle cover to 20%. Honeysuckle percent cover for all test plots was lower than the untreated control plot, which was around 80% honeysuckle cover.

Figure 22 shows that goat grazing in 2025 did not have a large impact on honeysuckle average height. Honeysuckle average height in 2025 was less than 2024 primarily due to mechanical clearing.

Because the tall honeysuckle had been mechanically cleared in 2024, resprouts and new seedlings were at goat graze height in 2025. The goats stripped the leaves from essentially all honeysuckle in the study area. Photos taken during the stem counts show the difference pre- and post-grazing in the study areas (Figures 23 through 26). The honeysuckle grew leaves in the spring, but after being grazed over the summer, there weren’t many new leaves by fall. Photos of the untreated control area A5 (Figure 27) show heavy honeysuckle foliage in the spring and somewhat less foliage in the fall, but still more honeysuckle foliage than the treated test units.

Discussion

Method comparison

All test plots had lower honeysuckle infestation at the end of the study compared to the untreated control plot. If the treatment sequence had a large impact on effectiveness, then one would expect the metrics at the end of the study to differ among test plots. In fact, Figures 20 through 22 show little difference among the test plots for data collected on September 29, 2025 at the end of the study, except for test plot A2b, which wasn’t mechanically cleared and therefore had more honeysuckle infestation than the other test plots. This reveals that mechanical clearing is required, but exactly when it occurs in the treatment sequence is not particularly important.

In 2024, two rounds of goat grazing were done, one in spring and one in fall, in units A1a, A1b, and A2b. The second round of goat grazing did not result in a measurable impact on the metrics at the end of the study. Unit A2a, which did not receive a second round of goat grazing by design in 2024, fared just as well as the units that were grazed twice in 2024. It would be better to save the expense of the second round of goat grazing in the first year and instead shift to mechanical clearing after the first round of goat grazing.

If one is grazing a sunny area, then a second round of goat grazing in the same year may be helpful, but this study did not have enough sunny areas to be conclusive in that regard. Unit A1/2e, the mulched edge, received more sun and grew back much quicker than in the shady forest, but unit A1/2e was a limited area and we aren’t able to draw firm conclusions about whether grazing A1/2e twice in a year was helpful.

Goat grazing in the woodland was a beneficial first treatment. IPRM personnel found it easier to operate the skid steer after goats had grazed, because operators can see where they are going. Otherwise, it’s a “wall of green and can be claustrophobic.” Also, bidding on a project is easier after goats have grazed, because contractors can see what they are bidding on.

To reduce honeysuckle resprouts, pulling honeysuckle out by the roots is preferred to cutting or mulching. A cut or mulched stem still has a mature root system and produces many sprouts off the cut stump.

Roots of pulled honeysuckle should be placed in the air to prevent re-rooting. Alternatively, the pulled honeysuckle can be piled and burned or mulched. Burning has the advantage of making the honeysuckle disappear, while a layer of mulch on the ground may help reduce invasive weed pressure and support a prescribed burn. More time is needed to assess whether burning or mulching honeysuckle turns out to be better.

Erosion impact

The impact of the various clearing methods on erosion is worth noting. Targeted goat grazing has low impact. Goats are not heavy, their hooves are small, and they are moved to a new portable paddock before they have time to create trails subject to erosion. Mechanical clearing with chain saws also has low impact. Workers may slip somewhat on the steep hillsides, but generally don’t cause observable erosion. A skid steer is very high impact. Skid steers are heavy and their track wheels tear up the soil and cause ruts and berms when turning. Scraping the ground with a bucket attachment to uproot small honeysuckle rips up the top layer of soil. Pulling large honeysuckle out by the roots leaves holes and disturbed soil. Pushing honeysuckle across the ground into piles disturbs the top layer of soil. Skid steers sometimes get stuck, leaving deep ruts. After clearing with a skid steer, the area has loose, bare soil prone to erosion. Therefore, I recommend clearing only once with a skid steer, with subsequent clearing done by goat grazing or human manual labor.

A skid steer is very effective at bush honeysuckle removal. A skilled operator can remove nearly all honeysuckle of all sizes by the roots in machine-accessible areas, whereas goats can’t kill honeysuckle they can’t reach and cutting with a chain saw results in bushy resprouts. The tradeoff for highly effective clearing of a skid steer is more erosion. Using a skid steer to mulch honeysuckle may help mitigate erosion risk by leaving a mulch layer on the ground after clearing.

Effect of seeding

Light seeding did not appear to have any advantage with regard to herbaceous groundcover. It should be noted that native grasses can take several years to get established, so perhaps in a few years a difference between seeded and unseeded areas will be apparent, but right now, there is no discernible difference in groundcover.

Prescribed burn preparedness

There is not a large difference among the test units with regard to preparedness for a prescribed burn, indicating the treatment sequence was inconsequential. There is still not enough herbaceous groundcover in any of the test units to carry a strong fire with enough residence time to adequately control bush honeysuckle. A higher fuel load of thatch, leaf litter, and larger fuels such as downed limbs or mulch is needed to increase heat residence time. Establishment of burnable groundcover may require removal of cull trees, because the forest floor is still shaded even after bush honeysuckle removal.

We plan a prescribed burn in winter/spring 2026 in the test areas and the untreated control area. We expect the burn to be spotty, burning best in patches of continuous leaf litter from the oak trees. There is limited, patchy mulch on the ground in unit A1b (see Table 3) but we don’t expect this to make a significant difference for the prescribed burn.

In February 2026, we observed that the untreated control area had wetter, more degraded leaf litter than the test units, probably due to heavy bush honeysuckle in the untreated control area. The test units, having been cleared of bush honeysuckle, have more sunlight and wind at the forest floor to dry out the leaf litter and therefore have a higher probability of ignition than the untreated control area.

Safety is always a top priority during a prescribed burn and the test units will be safer to burn than the untreated control area. Figure 28 shows representative photos of the untreated control area A5, in which heavy underbrush precludes easy access, compared to test unit A1b, which is more open. In unit A1b, the fire crew will be able to walk through more readily, see each other, and more easily communicate with each other, thereby improving safety of the burn.

Compared with the untreated control area, the greater accessibility of the test units enables ignition of the interior as well as the edges for a more complete burn. In contrast, it would be difficult to ignite the relatively inaccessible interior of the untreated control area.

In general, it can be hard to find an appropriate burn window. Because the untreated control area is unlikely to burn well, we will try to burn it when the sun is at its highest point in the sky so more sun is reaching the ground at the time of the burn. The fact that the test units are expected to burn better gives us a longer time window for burning the test units, not just when the sun is at its highest point in the sky.

Conclusions

The recommended method for controlling a severe infestation of bush honeysuckle without herbicide is a treatment sequence of (1) goat grazing in the spring to open up the understory for visibility, (2) heavy mechanical clearing in the fall to remove the honeysuckle by the roots as much as possible, especially the tall honeysuckle that goats can’t reach, and (3) goat grazing the following year to knock back resprouts and new seedlings.

For the prescribed burn planned for winter/spring 2026, the test units are expected to burn better than the untreated control area due to dryer, more intact leaf litter in the test units. Neither the treatment sequence nor light seeding of native grasses appeared to have a large impact on preparedness for a prescribed burn; burnability is more a function of the presence of continuous dry leaf litter on the ground.

I anticipate that goat grazing will be needed for several more years in the study area to control resprouts and new seedlings. Once the seed bank is depleted of honeysuckle seeds, perhaps grazing won’t be as necessary and prescribed burning may be a sufficient management tool. This remains to be seen.

I have already applied results from this study to more acreage in my woodland. In 2025, I grazed area A4 and then IPRM mechanically cleared it, pulling honeysuckle out by the roots. I plan to graze area A4 again in 2026.

Kitchens, Nathan. September 4, 2024. Looking back at meteorological summer 2024. CIProud.com. https://www.centralillinoisproud.com/weather-headlines/looking-back-at-meteorological-summer-2024/

Rathfon, R.A., S.M. Greenler, and M.A. Jenkins. 2021. Effects of prescribed grazing by goats on non-native invasive shrubs and native plant species in a mixed-hardwood forest. Restoration Ecology 29(4): e13361.