Effects of Herbicide Application Following Mechanical Site Preparation on Long-term Forest Development: An Experiment with Picea glauca

Part of the Young Naturalist Awards Curriculum Collection.

by Allister, Grade 10, Ontario - 2009 YNA Winner


Reforestation in Canada has historically involved the use of herbicides for quick regeneration of crop species. However, the long-term effects of herbicides have not been studied in detail. The study examines the effects of herbicide application on long-term forest stand development. It is a re-evaluation of Picea glauca (white spruce) trees planted and sprayed with herbicides 2,4-D and 2,4,5-T esters in the Algoma region of Ontario (47°N, 84°W), in 1968, following mechanical site preparation by angle-dozer blade and shark-fin drum.

The study established by Haig and Curtis in 1974 was re-evaluated by measuring the growth of planted white spruce trees by height, diameter, and crown position. In addition, the surrounding unplanted tree vegetation was measured, noting species, height, diameter, and crown position, within 20-square-meter circular plots. From these measurements, basal area and biomass values were determined to determine the relative growth of white spruce, the effect on the surrounding trees, and the overall effect on the development of the stand as a whole between treated and control plots for both methods of mechanical site preparation. These results come 41 growing seasons after the original planting of the white spruce.

Individual white spruce sprayed with herbicide displayed higher growth and greater productivity (greater height, diameter, basal area, and biomass, and a more dominant crown position) but a lower survival rate and a relatively equal total plot biomass compared to control trees on the same site preparation. The productivity of the surrounding unplanted tree vegetation was found to be lower with the use of herbicides, and the total productivity output of the stands was found to be lower with herbicide application.


Living in Sault Ste. Marie, Canada, I have seen our forests exploited for commercial use for many years and have feared for the destruction of the natural habitat. I simply could not imagine our country, or continent, without its vast expanses of primeval forests, which are home to thousands of organisms as well as water-purification systems and recreational areas for humans. My concerns lead to a study on the long-term effects of herbicides in reforestation.


Lower (1938) explains that in the late 19th century, intense logging of Pinus strobus (white pine), Pinus resinosa (red pine), and Picea glauca (white spruce) initiated concerns about forest conservation in Ontario. Armson (2001) discusses how the forests continued to be overharvested and were often unable to regenerate naturally “because heavy slash and residuals were often left after harvesting” (p. 12). This problem spurred government interest in developing regeneration practices.

MacKay (1985) describes how many reforestation methods were tested and numerous short-term studies were carried out to investigate options for establishing the principal conifer tree species. However, short-term studies were often used, suggesting a need for long-term investigations of regeneration treatments on forest development.

In 1969, an herbicide study was superimposed on a portion of a larger experiment in a cut-over mixed-wood forest in Goulais, Ontario. The study involved two different forms of mechanical site preparation, with and without chemical site tending. Haig and Curtis (1974) reported the fifth-year results of the study. Wood and Dominy (1988) subsequently reassessed all the trees in 1985.

Research Question

How does herbicide application following mechanical site preparation affect long-term forest stand development?

Wood and Dominy’s (1988) plot diagrams and data sheets (obtained from the Great Lakes Forest Centre) were used to locate and re-evaluate the planted white spruce. I measured the height, diameter, and crown position of all the planted trees. In addition, using the 20-meter-square circular subplots previously installed by Wood and Dominy, I identified the species and measured the height, diameter, and crown position of all the surrounding trees, an aspect omitted by previous researchers. Calculations of mean tree height, diameter, crown position, basal area, and biomass were carried out for both planted crop species and the surrounding unplanted vegetation. The data obtained is 22 years from the last measurement, and 41 growing seasons after planting. My reassessment of past work and my inclusion of new additional data and information will be used to determine the effects of herbicide application on long-term forest development on two different types of mechanical site preparation.

Previous Studies

Allister at Site I.

The primary focus of the Haig and Curtis study (1974) was to determine the relative efficiency of four kinds of site preparation, two forms mechanical and two forms chemical. The two types of site were scarified by either an angle-dozer blade (referred to as blade) or shark-fin drums (drum). These processes of mechanical site preparation are referred to as scarification. Three plots within the block scarified by angle-dozer blade and four within the block scarified by shark-fin drums received herbicide spraying, while three blade and two drum plots were used as unsprayed control areas, respectively.

1969 Goulais River Valley Experiment Layout shows the mechanical site preparation and chemical tending treatments used within each plot. The maps represent arial views of the site preparation blocks 29 and 31.

After five years, Haig and Curtis concluded that site preparation with an angle-dozer without subsequent herbicide application produced the best results. Reduced competition, caused by the severity of the mechanical site preparation, offered high survival to seedlings. The next best results were obtained by site preparation with a shark-fin drum and herbicides, because herbicide spraying meant higher growth, and then without herbicidal spraying. The worst results were found with site preparation by the angle-dozer blade and an application of herbicides. The herbicides were not necessary, given the effective mechanical site preparation, and they were toxic to the planted spruce.

1969 Goulais River Valley Experiment Setup

A 1985 re-evaluation by Wood and Dominy (1988) examined the development of the white spruce compared to the competing woody vegetation in order to determine crop tree response to herbicide application. The report found that the application of herbicides “at an early age can effectively release conifers from competing tree and shrub vegetation” (p. 180). Trees in the areas sprayed with herbicides were generally larger than those in areas not sprayed, except for the trees in the scarified, angle-dozed block.

1969 Goulais River Valley Experiment Layout shows the mechanical site preparation and chemical tending treatments used within each plot. The maps represent arial views of the site preparation blocks 29 and 31.


Background Research

Sutton (1985) presents the rationale behind site preparation for vegetation management. Site preparation, “the treatment of a site prior to seeding or planting in order to facilitate the regeneration of that site” (p. 11), is specific to a certain crop tree species, and aims to reduce competition from competing species. Mechanical site preparation usually involves methods of disturbing the soil through mechanical use (i.e., by a tractor with a blade) and is often combined with chemical treating.

Campbell, Wood, Thompson and Iskra (2001) address the necessity for chemical site preparation. The purpose of chemical site preparation is to “minimize the negative effects of competing vegetation on crop trees that are seeded, planted or naturally regenerated after treatment” (p. 231). Bovey and Young (1980) discuss the principal herbicides used in forest management: 2,4-D (C8H6Cl2O3), 2,4,5-T (C8H5Cl3O3), and silvex. These herbicides are important in maintaining productive stands: “2,4,5-T is the preferred foliar treatment. Most other herbicides are toxic to pines, including 2,4-D.”


Forest stand development following reforestation treatments involves three principal components: the growth of the planted crop trees, the growth of surrounding trees, and the overall productivity of the stand. Within these three areas, four testable hypotheses regarding the long-term effects of herbicide application are examined. If white spruce seedlings are planted following mechanical site preparation, whether by scraping the soil with an angle dozer blade or creating narrow trenches with shark-fin drums, then an herbicide application will:

  • improve the long-term growth of individual planted trees,
  • increase the overall survival and productivity of the planted trees,
  • reduce the productivity of the surrounding unplanted trees, and
  • not affect overall stand productivity.

Herbicide application, by giving planted trees the advantage of limited competition, will allow the trees to continue to grow with this advantage. The planted trees will have improved long-term growth.

Given the competitive advantage for the planted white spruce, I expect that their long-term survival will be higher than in the plots without herbicide. The trees will face less competition and therefore have a greater chance of survival in the long term.

The initial release of conifers from competing vegetation analyzed in the short and medium-term studies (Haig and Curtis 1974, Wood and Dominy 1985) will not only be beneficial to the planted trees but also detrimental to the surrounding unplanted vegetation. The planted trees are given an early start and are able to acquire more resources, which in turn severely hampers the surrounding vegetation from developing.

On a stand basis, overall productivity will not differ between the treated and the control sites. The resource output of a forest is not decided by the success or failure of a single species. According to Burger (1994), “the maximum production attainable on a site, given unlimited time, is the site’s carrying capacity, which is controlled by abiotic factors” (p. 17).


Independent Variable: Chemical weed control; application of herbicides or no application of herbicides (control)

Controlled Variables: Climate, regional insect exposure, mechanical site preparation by angle-dozer blade or shark-fin drum, soil moisture, soil nutrients, soil texture and soil depth

Dependent Variables

Hypothesis 1— Planted crop species, white spruce individual tree growth: tree height, diameter, crown position, basal area, and biomass

Hypothesis 2- Planted crop species, white spruce survival rate and stand-level productivity (basal area, biomass)

Hypothesis 3- Stand-level productivity (density, basal area, biomass) of planted and unplanted trees

Hypothesis 4- Overall stand productivity (total basal area, total biomass)


  • diameter tape (accurate to 0.1 centimeters)
  • Haglöf Sweden ® Vertex III Hypsometer Electronic Height Surveyor (receiver and transmitter), accurate to 1%
  • 2.5-meter measuring stick
  • 1.8-meter measuring stick
  • 1.3-meter measuring stick


A young man, Allister, the Young Naturalist award recipient standing in a forest wearing an orange safety vest and dark cap.
Allister at Site II.

Necessary equipment was gathered at Whitman Dam Road site (47°N, 84°W).

The map was consulted and the first plot in Block 29 was located. By starting at the beginning of the plot (labeled on Figures 1 and 2 with an ‘s’), the first stake was found. The tag was checked to ensure correct plot location.

In examining Wood and Dominy’s data sheets, the trees expected to be living in the plot (the trees measured during the last study and therefore alive) were determined. These were the trees that needed to be found in order to be re-evaluated.

Following the two-row system, I found the first tagged white spruce tree. The tags were located in the ground at the base of the tree; a sample tag is shown at the right. I recorded the condition of the tree (alive or dead).

By using the 1.3-meter measuring stick, the 1.3-meter height on the tree trunk was determined. The diameter tape was then used to determine the diameter of the tree at this height. The measurement was recorded.

The 1.8-meter measuring stick was used to determine a 1.8-meter mark on the tree. The Haglöf Vertex III transmitter was attached to the tree bark using the pick on the rear of the device.

This is a labeled stake, found at Block 29, Plot 1, Subplot 6.

I backed up approximately 20 meters from the tree (or to the closest position where both the electronic height surveyor transmitter and the crown of the tree were in view), and determined the height of tree by:

  • viewing the transmitter through scope of receiver and placing the cross-arm directly on transmitter node;
  • holding the receiver button until it registered in the system with an audible click;
  • pointing the receiver upward, resting the scope in line with the crown, and holding the receiver button;
  • repeating steps 8a to 8c three times, with mean height results displayed at end of each interval, and recording the measurements.

While in this current position, the crown position of the tree was visually evaluated on a six-point scale (Appendix 1). I recorded these measurements.

Steps 5-9 were repeated for all the trees in the plot. If a dead tree was found, it was recorded as such and only its diameter was measured.

I located the stake marking the start of the plot once again. The 2.5-meter plot stick was then rotated around the stake in a circular fashion, making the stake the center of the arc. I recorded each planted white spruce that was in the circular area as being “in” the circular plot.

For trees within the plot that were not tallied in the original study (the surrounding unplanted trees), the species of tree was determined and recorded. Steps 5-8 were repeated for all of these trees and Step 11 for remaining five subplot stakes.

I repeated Steps 2 to 11 for the remaining plots in Block 29.

I repeated Steps 2 to 13 for Block 31.


Part I

In terms of the long-term growth differences between individual white spruce trees, the effect of herbicide application is quite noticeable. Figures 3 and 4 depict the mean values for planted tree diameter and height, respectively, for the two types of mechanical site preparation and herbicide application. The highest mean tree heights, 10.8 meters and 8.9 meters, were found in the treated portion of the block, and the same trend continues for diameter (12.3 centimeters and 12.2 centimeters). It is evident that for each type of mechanical site preparation, the treated trees demonstrated greater growth, shown again in Figure 5 by the lower mean for crown position. This higher growth translated into more productive trees in terms of biomass and basal area. In the treated plots, average tree biomass (52.6 kilograms, 56.0 kilograms) and average tree basal area (130.9 cm2 and 141.5 cm2) with drum and blade treatment, respectively, were higher than in the control plots.

Table 2. Crop Species: Planted White Spruce Long-Term Growth Means.
Table 3. Crop Species: Planted White Spruce Survival Rates.

Given these results, the hypothesis is accepted. Herbicide-applied trees displayed larger growth: taller heights, greater widths, and more dominant crown positions.

Table 4. Mean Plot Biomass Totals.

Herbicide spraying, as discussed by Wood and Dominy (1988), can effectively release conifers from competing vegetation, and this trend is shown to continue in the long term. The advantage given to spraying was significant enough to allow for enduring effects; once in a dominant role, the planted trees were able to maintain this position and accelerate their growth over surrounding vegetation. Without herbicide, planted trees in both blocks were smaller in all aspects, lacking the competitive advantage.

Table 5. Mean Plot Basal Area Totals.

Height results differed from those found in 1985. While the angle-dozed block maintained the trend of higher tree heights in the plots with herbicide, the drum-prepared block did not experience substantial differences in height resulting from herbicide application. My 2007 measurements show that the average height for the control trees was 9.2 meters, compared to the herbicide-treated value of 10.8 meters. This data suggests the necessity of long-term studies relating to the full extent of herbicide effects.

Figures 3-6.

Despite the consistency of these results with those found previously regarding the growth effects of herbicide application, the findings still offer considerable merit. Treated trees can maintain a considerable advantage in the long term. Reforested stands with herbicide spray have crop species with higher growth development in the long term.

Part II

Figure 8 illustrates an unexpected trend: The control trees in both mechanical site preparations had a higher rate of survival; herbicide-treated trees experienced significant lower survival rates. In the drum-prepared block, the mean survival rate for treated trees was 43% and for control trees 53%, a moderate difference. For the blade-prepared block, mean survival was 70% in the control trees and 51% with herbicide application, a sizeable difference.

Figures 7-10.

The hypothesized higher survival rate of trees in plots treated with herbicide compared to the control plots was

Figure 11.

not observed in this study. The hypothesis in this instance is refuted.

Reduced competition from surrounding vegetation for the planted trees did not give the treated trees a higher survival rate. Lower survival in the treated plots can be attributed to two causes. First, the herbicide 2,4-D is actually toxic to conifers, and given the effective site-preparation procedures (prescribed burn, mechanical scarifying), much of the vegetation was already subdued, and thus the white spruce seedlings were affected by the excessive amount of herbicide sprayed around them. Second, the lack of forest canopy or surrounding trees exposed the treated trees to the harshness of the northern Ontario weather, especially in the winter. Trees in the control plots not sprayed by toxic chemicals were more protected from the elements by the surrounding vegetation, allowing for higher survival rates.

Average Plot Total: Plotted Basal Area vs. Surrounding Basal Area
Average Plot Total: Plotted Biomass vs. Surrounding Biomass

These results are highly valuable to a forest manager. Herbicides will improve the productivity of individual crop trees, as identified earlier, but also reduce the number of trees that survive. In terms of long-term forest development, herbicide use offers a stand with dominant trees and lots of growth, but fewer trees present than would occur without herbicide spraying.

Part III

The relationship between the productivity of planted trees and the surrounding unplanted trees is vividly pictured in the basal area and biomass charts, Figures 10 and 11. Both graphs clearly show that the controlled plots are composed of an almost equal proportion of planted and unplanted values (in basal area and biomass). However, the treated plots experienced a substantial decline in these values in the unplanted trees. In terms of biomass, using a drum-prepared block made little difference in the planted tree values (570.1 kilograms and 649.3 kilograms) of the treated and the control sites, while the unplanted tree biomass values fell, to 602.3 kg in treated plots from 898.2 kg in the control plots. This same trend held in the angle-dozed block (533.7 kg for control plot biomass and 455.1 kg for treated, 182.9 kg for control plot biomass and 602.3 kg for treated), and again for the basal area of both mechanical site preparations.

The hypothesis is accepted. Herbicide application had a very significant effect, lowering the productivity of the unplanted surrounding trees considerably. The herbicide was detrimental to the productivity of the surrounding vegetation. Herbicides effectively killed the competing trees, and the sprayed conifers, upon being released from competition, were allowed to develop at the expense of the unplanted trees.

However, it is interesting to note that total planted tree basal area and biomass on a stand level did not differ greatly between the treated and control sites for each site-preparation method; the major reduction in biomass and basal area occurred in the unplanted trees within the treated plots.

Part IV

Site III with Allister in background.

The total productivity of each site is also demonstrated in Figures 10 and 11. Total biomass in the drum-prepared block was much higher in the control plots (1,547.5 kg vs. 803.3 kg) and in the blade-prepared block (1135.9 kg vs. 638.0 kg). This trend continued for the basal area values: 4,459.1 cm2 in the drum-prepared control and 2,164.3 cm2 in the treated, 3,631.0 cm2 in the blade-prepared control and 1,619.6 cm2 in the treated.

The hypothesis in this instance is refuted. Total productivity between the control and the treated sites was expected to be relatively equal, a product of the abiotic factors of the site and the initial site preparation. However, the drastic productivity changes between the treated and the control sites shows herbicide application does affect overall productivity.

The success of the planted trees was expected to counterbalance any shortcomings in the growth of the unplanted trees in the herbicide-treated plots, which would mean the total biomass of the plot would be equal to the control site. However, the herbicides were very effective in killing the surrounding vegetation when spraying occurred, and the advantage given to the planted white spruce inhibited the regrowth of surrounding vegetation to the site’s full potential. (The site potential is demonstrated in the control plots.)

Figure 9 offers a unique perspective on the biomass data. The graph shows that the biomass of the planted trees is essentially the same, regardless of whether mechanical or chemical site preparation was used. This suggests that the low average survival rate per plot (Part II) of treated planted trees and the high growth experienced by these same treated trees (Part I) cancel out in terms of productivity; both processes create relatively similar final biomasses.


The results of this study demonstrate that the application of herbicides at an early age, which frees the tree crop from competition, has a substantial impact on the development of a forest stand. Two-year-old seedlings maintained higher growth and productivity than trees with identical site preparation for 41 growing seasons, having a higher mean height, diameter, biomass and basal area, and a more dominant crown position. Herbicides lower the survival rate of the planted species because of their toxicity and by preventing any kind of vegetative cover from the weather for seedlings. The surrounding unplanted vegetation was affected negatively by herbicide spraying. Herbicides lowered the productivity of the surrounding vegetation because the planted crop species dominated, with higher crown positions and a larger basal area, thus capitalizing on available resources. The long-term development of a forest in its total productive output is affected substantially by herbicides and is not controlled entirely by the resources available.

The results of this study imply that although herbicide spraying can improve the growth and productivity of selected crop species, it does not increase the biomass of the crop species and has a negative impact on the survival of that crop and the potential productivity of the site as a whole, regardless of the mechanical site preparation. Such a result clearly calls the need of herbicide spraying into question, while the different results in immediate, midterm, and long-term studies reinforce the necessity for continued long-term studies on reforestation practices.


Armson, K.A., W.R. Grinnell, and F.C. Robinson. “History of reforestation in Ontario.” In: R.G. Wagner and S.J. Colombo, eds. Regenerating the Canadian Forest: Principles and Practices for Ontario. Markham, Ontario: Fitzhenry   Whiteside, 2001.

Bovey, Rodney W., and Alvin L. Young. The Science of 2,4,5-T and Associated Phenoxy Herbicides. New York, New York: John Wiley   Sons, 1980.

Campbell, R.A., J.E. Wood, D.G. Thompson, E. Iskra. “Site Preparation: Chemical.” In: R.G. Wagner and S.J. Colombo, eds. Regenerating the Canadian Forest: Principles and Practices for Ontario. Markham, Ontario: Fitzhenry   Whiteside, 2001.

Haig, R.A. Operational Trials of Site Preparation and Planting Methods in the Goulais River Area Ontario. Sault Ste. Marie, Ontario: Canadian Forestry Service, 1969.

Haig, R.A., and F.W. Curtis. Cost-Effectiveness of Four Methods of Establishing White Spruce on a Cut-Over Mixed-Wood Site in the Goulais River Area, Ontario. Sault Ste. Marie, Ontario: Canadian Forestry Service, 1974.

Korzukhin, Michael D., and M.T. Ter-Mikaelian. “Biomass equations for 65 North American tree species.” Forest Ecology and Management 126 (1997): 1-24.

Lower, A.R. The North American Assault on the Canadian Forest. Toronto, Ontario: Ryerson Press, 1938.

MacKay, D. Heritage Lost: The Crisis in Canada’s Forests. Toronto, Ontario: MacMillan of Canada, 1985.

Sisam, J.W.B. “Forest Development on the Goulais Watershed, 1910-1933.” Sault Ste. Marie, Canada, Department of Mines and Resources, Silvicultural Research Note 55, 1939.

Wagner, R.G., and S.J. Colombo. Regenerating the Canadian Forest: Principles and Practices for Ontario. Markham, Ontario: Fitzhenry and Whiteside Ltd, 2001.

Weed Science Society of America. Herbicide Handbook: Fifth Edition. Champaign, Illinois: Weed Science Society of America, 1983.

Wood, J.E., and S.W.J. Dominy. ”Mechanical Site Preparation and Early Chemical Tending in White Spruce: 19-Year Results.” The Forestry Chronicle June 1988: 177-181.

Appendix I: Scientific Method Evaluation

The first limitation of this study is its narrow focus within the complex, interconnected properties of forests. The study examined only the health of planted and unplanted trees within a circular plot. The study does not examine many other facets that are responsible for the health of these trees besides herbicide spraying and site preparation, including topography, soil nutrient analysis, animal and pest disturbance, and historic climatic conditions. As such, although an excellent basis for a long-term study on herbicides, examinations of these variables and their connections is essential for accurate depictions of forest stand development.

Another, related limitation is the issue of controlling all the variables. The experiment site is located in a real forest, which is susceptible to weather, soil, climatic variances, pest invasion, and human contact. The reality of the testing site means some trees may be affected differently by uncontrolled variables than others. It is not possible to control the long-term growth of trees, as is done in a laboratory or greenhouse for experiments with short-lived plants or seedlings.

Additional limitations to my scientific study involve sampling techniques. Crown position was observed visually, height was determined by tilting the surveyor to the angle of treetop, and measuring each tree diameter at exactly 1.3 meters was essentially impossible due to branches and uneven terrain. These three techniques rely heavily on the observer, and as such their accuracy is limited by human error.

The sites used within my report offer several avenues for future research and improvement of the study’s limitations. Given their history, these sites will be very effective for continued long-term studies on the effect of herbicides on tree and stand growth, mechanical site preparation, and their interaction. In addition, the effects of herbicides on other environmental components, such as soil nutrients, minor shrub vegetation, and animal habitat, can be examined.