Works Best On:
Forbs, seedlings
Cost (acres/year):
Moderate ($500–$1000)
High risk
Moderate risk
Cultural Resources:
Low risk
Environmental Hazards:
Low–moderate (erosion, ground-dwelling animals)


Flaming is used to control carpets of seedlings of broadleaf weed species that germinate together in large numbers and are localized in small areas. It typically involves sweeping a propane torch flame over very young plants, causing their cells to rupture from the heat. Flaming is most effective on seedlings from the cotyledon stage to the six true-leaf stage. The ability of this technique to outpace production of propagules is dependent on implementation of technique and timing.

Flaming was originally developed as a weed control technique by the agricultural industry to treat weeds in furrows using a tractor and a boom. It has only recently been modified to be used as a hand-held weed control technique for wildland and urban landscape settings. This method has a narrow range of conditions under which it can be applied effectively, including high site moisture (too wet to support fire), small spatial scale, accessibility to roads, and early phenological stage of plants. Despite this narrow range, it is a helpful tool to incorporate into an IPM program because it lengthens the weed treatment season by effectively targeting seedlings early in the season that would otherwise take more time to control manually. Flaming also has the advantages of not leaving plant debris behind and leaving soil surfaces intact. Flaming may have the added benefit of forcing germination of seeds in the soil. With follow-up treatment, this technique can shorten the longevity of the seedbank.

Flaming is intended to wilt a plant, not burn it. Burning with a propane torch takes significantly more time per plant than flaming, is not any more effective, and poses a greater risk of fire. Seedlings that have their growing tips above-ground and have poorly developed roots are the most susceptible. Grasses and other monocotyledonous plants with their growing tips at or below-ground or at ground level are likely to regrow after flaming unless they are very young seedlings. Older plants with well-established root systems cannot be effectively controlled with this technique unless they have no capacity to regrow after being damaged. Plants that reproduce vegetatively cannot be controlled with flaming unless repeated treatments are used to deplete carbohydrates stored in roots below ground. This technique has been used effectively for seedlings of broom species, poison hemlock, stinkwort, yellow star thistle, and very young grasses. Broom seedlings in particular are highly susceptible when they are in the 3-4-leaf stage and younger and germinating synchronously.

Flaming can have non-target effects such as killing desirable plants in its path that are unable to regrow. Recruitment of desirable competitive species may be an important consideration in your long term weed management plan and should be considered. Alternatively, there may be very little desirables in the seed bank if the weed infestation is old enough to have exhausted other species. In this instance there may be very little desirable species that can immediately recruit. Additionally, flaming has the potential to kill any ground-dwelling organisms immediately in the path of the flame.

Flaming should not be conducted in dry conditions or with excessive thatch that might catch fire. Fire extinguishing equipment (water and fire rakes) should always be on hand in the event that vegetation catches fire. Flaming is not generally recommended for seed bank control because of the high heat required to kill seeds. It is most efficient for small areas because it is both very labor- and fuel-intensive at larger scales.

This method is moderately expensive both because of the cost of propane and the labor to do the flaming and refill tanks. Furthermore, applicators need to incorporate down time if equipment freezes from prolonged use (“frost-up” due to gas expansion). Some areas will require site preparation to remove flammable debris.

Some practitioners also use propane torches to burn individual plants in order to kill their seeds at maturity before they are released. This late-season approach targets seed rather than growing plants and has been used effectively on barbed goatgrass. However, the risk to applicator and of fire spread is much greater with burning, especially late season and should only be conducted with proper permissions, sufficient training, and appropriate protective equipment. It is not specifically recommended here.

How to Use

There are a variety of torches that produce different flame intensities. Weed control can be effective with 400,000 - 750,000 BTU burners. Several different cylinder (tank) sizes are also available. Non-agricultural applicators usually use a stand-alone 20lb or higher propane cylinder with a hand-held torch attached with a hose. Larger/heavier cylinders can be used but need to be transported by ATV or dolly. Small-sized backpack propane tanks are also available. Hoses can be connected together to increase accessibility. Note that tanks, hoses, and orifices can freeze up during use, so warmup periods or back-up tank/torch set-ups should be planned for. Use quick release air couplers to trade tanks quickly or use larger tanks to reduce frost-up. Torch mounts are also available for vehicles and ATVs for larger areas, but may have limited utility in a wildland setting.

Cylinders may be steel or fiberglass. Fiberglass tanks are lighter and generally considered safer. Torch arrays should include a POL valve to protect a tank from over-filling. All valves should be tightened securely. Torch wands should be ergonomic and light enough to minimize applicator strain. They should have a standing pilot light. Consider also using a torch with piezo-electric self-ignition for convenience. Prepare to have a sparking igniter if the built-in ignitor fails.

Before flaming, rake away organic debris to make flaming more efficient, to reduce risk of fire, and to avoid killing insects, reptiles, and amphibians. Perform a test burn to ensure that the flame will not carry in the vegetation where you will be flaming, To perform a test burn, as illustrated by Ken Moore (see flaming video link), scrape a small test plot in an area that is representative of your site. Scrape all vegetation off to mineral soil. Flame the vegetation in this burn plot. If the vegetation carries flame to the mineral soil line, this indicates that your site is too dry to use this technique at this time without risk of fire. For best results, plants need to have enough moisture for heat to burst tissue cells and cause wilting. Treated plants will become discolored and will bend or collapse. Ideal conditions are when target plants are moist enough to be scalded, rather than burned, by treatment. Even if you may have sufficient moisture in your test site, be cautious about gusting winds that may move burning materials outside of this area, potentially starting a fire. Do not use this technique if wind is strong enough to dislodge debris.

Proper flaming technique involves passing the torch flame over seedlings in a slow, even motion. If plants have sufficient moisture, they will wilt as a result of being steamed, rather than burned. In rare situations (i.e., when conditions are safe, plants are susceptible, but too late in their growth stage to be vulnerable to steaming), plants will be singed instead, though this is less efficient and more risky, and therefore not preferred.

Flaming can stimulate seed germination. If it does, it will require follow-up treatments later in the same season. For many fabaceous species (e.g., broom and acacia), multiple treatments will be necessary throughout mid-winter and early spring each year in order to ensure that most germinating seedlings are controlled. This technique can be used to flush the long-lived seed bank of some invasive species. Its lower efficiency may be offset by a reduction in the number of repeated annual treatments if the weed seed bank can be exhausted more quickly. Flaming must be conducted repeatedly over time until the weed seedbank is depleted in order for this tool to be effective over the long-term.

Special Tips

Species with highly synchronized seedling flushes are ideal for flaming because seedlings will all be of a similar age.

If weeds are flamed on steep or rocky terrain, tanks will need to be secured either by being mounted on a backpack or by otherwise being held securely to avoid rolling away.

Many practitioners recommend spraying seedlings with water in dry conditions ahead of time to increase efficacy by increasing heat conductance.

Flaming, though not very selective, can be fairly precise, so it can be used to kill small seedlings around already established plants to maintain desirable plant cover and minimize soil disturbance.

The valves on the propane tank will freeze up after continuous use. It is recommended to have two tanks and alternate their use. Do not use the flame to thaw tanks. Also, when tanks freeze, dirt and mud with weed propagules can freeze to the bottom if it is contact with the ground. Be careful to clean the tank before moving or cradle the tank in a crate or other carrying device to avoid contact with the ground.

Optimal Conditions for Use

Moist conditions, low thatch cover, low to no wind, and seedlings at or younger than the 6-leaf stage (or less than 2 inch diameter rosettes).


Because this technique uses fire, it can cause concern with some members of the public, as well as land managers. Ensure proper safety protocols are being followed and considering installing a temporary interpretive sign describing treatment to the public.

Flaming around poison oak or its litter can cause severe complications for people who are sensitive to this species. Under these conditions, applicators should consider being fitted for and wearing a respirator. Perennial grasses may be favored by flaming and burning because these techniques stimulate tillering and spread.

Thistles are not effectively treated by flaming. Even when still very young, they have deep tap roots and require 2+ treatments to die, making other methods more effective and efficient. Young shrubs and trees that have deeper roots may also resprout after initial wilting.

Flaming is difficult to use on summer germinating species because of fire danger under dry summer conditions.

When used over bare ground, flaming maintains bare surfaces but can also increase the risk of erosion.

Flaming is equally effective for low-density invasive plant stand as for high-density stands but is more inefficient. Consider alternate techniques such as hand pulling or scraping for less dense stands.

Flaming can negatively affect desirable species in the treatment area. Identification of desirable plant species, angle of the flame, intensity and timing can mitigate these off-target effects and increase selectivity.

Potential Hazards to Humans, Environment, and Cultural Resources

Human safety. Medium risk. Skin burns are a risk for applicators. Clothing can catch fire. Flaming also creates fumes and smoke that can be inhaled.

Cultural resources. Low risk. Cultural resources are at low risk because this technique does not disturb the soil surface.

Sensitive species. Low to medium risk. Flaming can kill insects, amphibians, and reptiles in the immediate treatment area. Applicators should flush a site prior to beginning treatment. Raking loose vegetation, duff, litter, etc. from the site will help achieve the same goal.

Habitat. Low to medium risk. Loss of vegetation, duff and organic material can negatively affect insects, amphibians and small mammals. This technique is best in small scales so that habitat elements are not removed or disrupted in the environment.

Erosion. Medium risk. Erosion may be a significant consideration since flaming treatment occurs during the rainy season and can result in bare ground. Take precautions to avoid erosion on steep slopes, near streams and other water bodies.

Fire. High risk. It is essential to perform flaming in moist conditions and with fire suppression tools nearby. Perform a burn test before beginning.

Consider Combining with the Following Non-Chemical Methods

This technique is most effective when followed up by manual removal of plants that emerged after treatment or that did not die as a result of flaming. Where additional weed seed is stimulated to germinate, follow-up flaming can further reduce the seed bank. Mowing, raking, or brush removal prior to flaming will reduce flammable thatch and increase surface area for seedlings to germinate.

Don’t Use This Technique When/For

Do not use flaming under high moderate or high fire risk conditions. Dry, windy conditions and sites with substantial dry thatch should be avoided. Avoid using flaming without prior surveys and employ avoidance measures if sensitive animals (arthropods or vertebrates) are known to inhabit a site. Consider materials for erosion control near streams, drainages, or waterbodies.

Supplementary Information

Some practitioners argue that weed species with higher seed production can be treated more effectively than those with lower seed production because they will typically germinate in large numbers and therefore be more apparent. This potential benefit should be weighed against the greater difficulty in being able to treat an entire population if it is already well established with a large seed bank.

Marshes are typically too wet to treat effectively because the saturated soil absorbs most of the heat from the torch. However, Limonium seedlings have been treated successfully in high marsh habitat. In this case, they were singed, rather than steamed. Flaming may be difficult and too risky to use in shrubland habitat because of the large amount of surface fuels present.

Although grasses are generally not effectively controlled by flaming, there are some exceptions. Flaming very early in the growing season, can reduce annual grass cover without harming co-occurring native annual species that haven’t germinated yet. Furthermore, one study had found that Japanese stiltgrass is effectively controlled over two successive years of flaming (see Ward and Mervosh 2012).

Whereas flaming is generally not considered effective on perennials, experiments in an agricultural setting suggest that some perennials could be reduced by >90% with three rounds of flaming. Grass control was not effective under these conditions. Timing of applications was critical in determining the level of efficacy. Propane use of 1.2 kg propane / km-1 was effective at reducing target broadleaf weeds early in their growth, but higher use of >2.6 kg propane/km-1 was less effective at later stages. High use of over 4.1 kg propane/km-1 was ineffective for grass species, regardless of their growth stage.

The drier the area, the more limited the window for optimal flaming becomes. Timing of germination and wet weather is critical to effective control with this technique. Extremely moist conditions lengthen the treatment timing, but also can make it less efficient as the ambient moisture absorbs BTUs, requiring more to treat target plants. In contrast to what most practitioners have found, Ulloa et al. (2012) conclude that efficacy is actually higher later in the day when relative leaf water content is lower. Flaming under dryer conditions should always be weighed against the increased risk of fire.


DiTomaso, J.M. and D.W. Johnson (eds.). 2006. The Use of Fire as a Tool for Controlling Invasive Plants. Cal-IPC Publication 2006-01. California Invasive Plant Council: Berkeley, CA. 56 pp.

Frey, M., Jenner Soong, J. Feeser, and S. Dishy. 2008. Identifying Control Techniques for Rumex acetosella in the Presidio of San Francisco (California). Ecological Restoration, 26: 109–111.

Hatcher, P.E., and B. Melander. 200). Combining physical, cultural and biological methods: prospects for integrated non-chemical weed management strategies. Weed Research, 43: 303.

Horesh, A., Y. Goldwasser, K. Igbariya, Z. Peleg, and R.N. Lati. 2019. Propane Flaming as a New Approach to Control Mediterranean Invasive Weeds. Agronomy 9: 187.

Moore, K. 2012. Flaming: A new tool for wildland weed control. Cal-IPC training video,

Niederer, C., et al. 2014. Identifying practical, small-scale disturbance to restore habitat for an endangered annual forb. California Fish & Game 100: 61–78. EBSCOhost,

Rask, A.M. and P. Kristoffersen. 2007. A review of non-chemical weed control on hard surfaces. Weed Research 47: 370-380.

Ward, J.S. and T.L. Mervosh. 2012. Nonchemical and herbicide treatments for management of Japanese stiltgrass (Microstegium vimineum). Invasive Plant Science and Management 5:9-19.

Ulloa, S.M., A. Datta, C. Bruening, G. Gogos, T.J. Arkebauer, and S.Z. Knezevic. 2012. Weed control and crop tolerance to propane flaming as influenced by the time of day. Crop Protection 31:1–7.

Authors and Credit

Lead Author: Pamela Beitz,Integrated Pest Management Specialist, East Bay Regional Parks District


David Thomson, Senior Ecologist, San Francisco Bay Bird Observatory
Pete Frye, Independent Contractor

Additional Contributors:

Ken Moore, Wildlands Restoration Team
Darren Bresee, Park Ranger, East Bay Regional Parks
Patrick McIntyre, Senior Ecologist, NatureServe


Ratings: Excellent (>95% control); Good (81–95% control); Fair (50–80% control); Poor or ineffective (<50% control)

Results are based on an estimation of maximum possible single-season reduction in weed cover and propagule production (=control). Control efficacy was scored for each plant and site characteristic for each management practice individually using best available information, assuming other conditions were optimal. Results for management practices are organized by efficacy rating based on the lowest rating they received for the combination of plant and site characteristics chosen. Rating results provided by the WeedCUT tool are generalized and may not be suitable for all plants or site conditions with the characteristics chosen. Ratings assume that a multi-year strategy will be employed to achieve management goals.
Plant Flowering Period
    Winter Good
    Spring Good
    Summer Good
    Fall Good
    Multiple Seasons Fair
    None Poor
Plant Germination
    Winter Good
    Winter / Spring Good
    Spring / Summer Poor
    Opportunistic Fair
Plant Palatability
    Yes No Information
    No No Information
    Partial No Information
Plant Plant Growth Form
    Grass Good
    Forb Good
    Shrub Fair
    Tree Poor
    Vine Poor
Plant Plant Type
    Annual Good
    Biennial Good
    Perennial Good
Plant Propagule Production
    Low (<1000/square meter) Good
    Moderate (1000–10,000/square meter) Fair
    High (>10,000/square meter) Poor
Plant Rate of Spread
    High (doubling in <10 year) Fair
    Moderate (50–75% increase in 10 years) Fair
    Slow Rate (25% increase in 10 years) Good
Plant Resprouting / Regenerative Capacity
    Low Good
    Moderate Poor
    High Poor
    None Good
Plant Seed Life
    Short (≤3 years) Good
    Moderate (4–10 years) Fair
    Long (>10 years) Poor
Plant Type of Reproduction
    Seed Good
    Vegetative Poor
    Seed & Vegetative Fair
Plant Type of Vegetative Reproduction
    Rhizome / Stolon / Stem Poor
    Bulb / Corm / Tuber Poor
    Root sprout / Sucker / Crown sprout Poor
Site Existing Desirable Plant Cover
    <10% Good
    10–25% Good
    26–50% Poor
    51–75% Poor
    >75% Poor
Site Ground Condition
    Muddy Good
    Smooth Good
    Cobbly Good
    Rocky Good
Site Habitat
    Marsh / Wetland Excellent
    Riparian Good
    Grassland Good
    Shrubland Good
    Woodland / Forest Good
Site Level of Tolerable Disturbance
    Low Good
    Medium Excellent
    High Excellent
Site Slope
    Flat Good
    Moderate (10–40%) Fair
    Steep (>40%) Fair
Site Target Area
    <40 square feet Excellent
    0.001–0.01 acre Excellent
    0.02–0.1 acre Good
    0.2–1 acre Poor
    2–10 acres Ineffective
    11–50 acres Ineffective
    51–100 acres Ineffective
    >100 acres Ineffective
Site Targeted Invasive Plant Cover
    <1% Good
    1–10% Good
    11–25% Good
    26–50% Good
    51–75% Excellent
    >75% Excellent
Site Vehicle Accessibility
    Roadside Excellent
    <100 feet from road Good
    100–1000 feet from road Good
    >1000 feet from road Poor