Burning Trash for Science: The Potential Use of Discarded Waste to Monitor Energy Fluxes Delivered to Ecosystem Components by Wildfires

Abstract

Assessing the energy flux delivered to ecosystem components by wildfires is hard because of technical and safety problems in performing measurements during such events. Here, we present a laboratory and field experimental assessment of a new method of evaluating a wildfire energy flux; our approach is based on the fact that different types of trash deform at different temperatures. We produced deformed trash in a laboratory environment using an iCone calorimeter to deliver a range of heat fluxes over a range of durations. We followed this by placing trash in instrumented prescribed fires. We show that finding melted or heat-altered plastic bottles and aluminium cans in the aftermath of wildfires can provide useful information about the heating that they received during the fire: plastic bottles are a useful indicator for areas that received less than 2 MJ/m² with a maximal temperature of <200 °C, while aluminium cans may be applied to higher-energy sites 100 MJ/m² that experienced a temperature above 600 °C. We provide a semi-quantitative proxy guide as to what different observed deformations may indicate in terms of energy flux and hope that this may allow scientists and forest managers to easily and cheaply assess the energy flux delivered to ecosystems and semi-quantitatively compare different wildfires.

1. Introduction

There is limited quantitative research regarding the energy flux delivered to ecosystem components by wildfires [1], largely because this is difficult to measure. Areas are not easily accessible during fires, making the installation of monitoring equipment challenging; hence, the majority of papers report fireline intensity, but instead of accessing it directly, they focus on easily measurable proxies such as flame length [2,3]. A more comprehensive picture of wildfire behaviour linked to measurements of the energy flux to ecosystem elements (e.g., tree boles) is obtained from relatively rare examples of instrumented experimental fires, e.g., [4,5,6]. In such cases thermocouples are often used to record temperatures they experience during the passing of a fireline and in response to post-fire front smouldering. Where thermocouples are properly calibrated, they can provide detailed, accurate and precise information on temperature in a specific location with a high temporal resolution. However, aside from the difficulty of undertaking full-scale experimental fires or positioning thermocouples ahead of wildfire flaming fronts, they are additionally relatively expensive and prone to damage during a fire. Because of this, only a limited number of devices are usually applied to describe large areas, which may not be representative for a highly heterogenous environment such as a natural wildfire [7,8]. Moreover, most experimental fires are usually performed outside of the fire season (i.e., within the prescribed fire season) and so are not fully representative of natural wildfires.

As a result, most in-season wildfires are described only by post-fire assessments from the surface [9,10] or orbit [11,12]. Therefore, thermal damage-linked post-fire effects on the environment and ecosystem elements are usually described only in a qualitative way [13,14,15]. For example, the descriptor of ‘Fire Severity’ is based on qualitative visual assessment of organic matter loss from an ecosystem following a fire [1]. Although this system may be useful in designing targeted post-fire mitigation strategies [16,17], forecasting level of survival of seedbanks [18] and post-fire recovery and succession [19], it is highly subjective and not consistent between different locations, ecological settings or individual assessors [20,21]. Therefore, the development of quantitative methods that capture the energy deposited by a wildfire and its influence on environment is crucial.

Currently available ground based quantitative methods include NIR spectroscopy and charcoal reflectance, while remote satellite-based options include difference normalized burn ratio (dNBR). NIR spectroscopy can be used to study the post-fire evolution of soil, but it requires special equipment and highly trained staff [21]. The reflective properties of charcoals collected after a wildfire relate to the total energy flux to the heated item/ecosystem component [6,18,22]. However, this method requires field sampling and long-duration laboratory processing and measurement, as well as a specialist microscope and an experienced operator, to be performed properly. Although those approaches may work well for research-intensive studies, they are not practical in most other situations. This challenge led us to look for other cheap and pragmatic approaches that could be used to easily and frequently gather thermal datasets from fires in any ecosystem type, with the aim of providing a basic quantitative estimation of the energy flux to ecosystem components during natural, prescribed and experimental wildfires.

Our aim is to provide a low-cost and rapid assessment approach that may allow ecologists to assess the energy distribution across a burned area by looking at the effects of vegetation-fueled fires on trash, such as tin cans, bottles and plastic items, which are often revealed following wildfire events. We suggest that thermally altered trash can serve as indicators of the amount of energy that thermally altered and deformed them because different types of materials, such as plastics, are known to deform at different temperatures. To develop a trash-based proxy, we surveyed the most common types of trash discarded in areas popular for outdoor semi-wild areas of recreational activity in the UK. Then, we performed a set of laboratory and field scale fire experiments to determine the energy levels required to cause certain levels of thermal deformation of the trash.

2. Materials and Methods

2.1. Trash Type

To establish the most common items of trash discarded in natural areas, we performed a series of field surveys within two fire-prone heathland locations in the Southwest of the UK that are commonly accessed by walkers, trail bikers, horse riders, etc. Different team members walked through different outdoor recreational areas, including the intersections of transport routes, paths and access points, as well as through more out-of-reach locations at Woodbury Common, Devon and Wareham Forest, Dorset. Each team member spent 4 h walking along the edges of footpaths, each covering 4 km. Approximately 2 m either side of the footpath through each of the two heathlands searched. The team members then each spent 1 h covering 1.5 km each searching alongside roads that bisected each heathland, again searching 2 m into the vegetation from the roadside. In all cases every piece of trash observed was collected and taken to the lab. By far the most common type of plastic waste found was that of drinks bottles, particularly Coca-Cola brand bottles; in many cases the bottles had their lids on. Most beverage bottles are made of polyethylene terephthalate (PET) that begins to soften at ~140 °C, melting > 260 °C and boiling > 350 °C. The second most common items were aluminium cans (melting point > 660 °C), the most common being Coca-Cola brand and various types of beer can. Based on these findings and to control for the influence of different material types, we performed all laboratory and field-based experiments using Coca-Cola brand bottles and cans.

2.2. Laboratory Experiments

Whilst the melting and boiling points for such common materials are available, the temperatures given assume equilibrium conditions, which are not applicable during a wildfire. Moreover, the response of an individual bottle type also depends on the thickness of the bottle walls and potentially their shape, which changes with the brand and beverage type. In general, heat flux measurements are more meaningful than temperature data; hence, we attempted to study how both heat flux and temperature regimes are linked to certain alterations in the material properties.

We performed a set of heating experiments by exposing standardised trash items to heat fluxes ranging from 10 to 65 kW/m² for 0.5 min for up to 40 min using an iCone calorimeter (Fire Testing Technology, East Grinstead, UK). Empty plastic bottles were exposed to lower heat fluxes of 10, 15, 20, 25 and, in a couple of experiments, 50 kW/m². Cans were investigated at heat fluxes of 30, 35, 40, 45, 50, 55 and 65 kW/m² for 20 and 40 min, because shorter exposure times did not produce any observable results. Each setting was run in triplicate.

During every experimental run (Figure 1), a single Coca-Cola brand bottle or can was placed in a metal baking tray where it sat on a 3–4 cm thick bed of sand and was exposed to the selected heat flux from above over the selected duration. The separation between the heat-source (conical heater of the iCone Calorimeter) and item was 25 mm. After the assigned time was up, the item was removed, wrapped in aluminium foil and covered with cold sand to smother any combustion reactions and cool it down, preventing any continued melting/combustion. For each heating experiment we recorded the temperature of the iCone heater element (°C), the resultant heat flux (kW/m²) and the temperature (°C) at the surface of the item being tested using a water-cooled heat flux meter (Fire Testing Technology, East Grinstead, UK) and a TC-08 Picolog temperature data logger, respectively (

Table 1. Cone heater set temperature, heat flux and mean temperature measured at the surface of the item being tested.

Cone Heater Set Temperature, °C	Measured Heat Flux at Item Surface, kW/m2	Temperature at the Surface of the Item, °C
413	10	75
482	15	110
547	20	135
600	25	165
645	30	180
681	35	205
715	40	225
750	45	260
778	50	285
806	55	301
830	60	323
855	65	358

). An additional set of experiments was undertaken with the cans, wherein a can was placed on a bed of Calluna vulgaris (Common Heather). This allowed us to increase the energy to which the can was subjected, where the fuel beneath also ignited. The Calluna vulgaris was found to have an average peak heat release rate of ~400 kW/m² and a total heat release between 25 and 35 MJ/m²; this meant that the burning heather beneath the cans plus the 50 kWm/² flux over 40 min totalled around 121–131 MJ/m² cumulative energy flux. All experiments were video recorded so that the timings of changes in the material state could be observed.

2.3. Field Experiments

Field experiments were undertaken during the UK prescribed fire season in collaboration with prescribed burns teams that comprised local landowners (Forestry England, National Trust and the RSPB) during winter 2020 in the southwest UK. Both fires were about 5000 m²: a burn near ‘Isolation Cottages’, at 50.658261° N, −2.082876° W, was undertaken on 29 January 2020 (later in the text referred as IC), and another burn near Furzebrook Road (later referred as FB), at 50.667366° N, −2.107124° W, was performed on 4 February 2020. During both fires it was sunny, ~10 °C and wind speeds were on average ~2 m/s, with some wind gusts up to 6–9 m/s. At the IC site, the vegetation consisted mostly of heather (mostly Erica cinerea and Calluna vulgaris) with Molinia caerulea in wetter areas (Figure 2a); the fire did not spread well, and areas of fuel mainly torched and went out on application of fire (Figure 2c). At the FB location there was a relatively heavy load of Ulex europeaus on the site, with an understory of heather (mostly E. cinerea and E. tetralix) and Molinia caerulea (Figure 2b); a head fire was allowed to spread rapidly across the site with long flame lengths (Figure 2d).

Each trash plot was instrumented with four type K-310 Stainless Steel Sheath 1.00 mm diameter  ×  1500 mm long thermocouples (supplied by TC direct, Uxbridge, UK) which were attached to Testo 176 T4 battery-powered LCD display temperature data loggers (Testo, Alton, UK). Twelve trash items were monitored at the FB site and forty-eight at the IC site. The thermocouples were placed into the opening of the Coca-Cola brand cans to keep them in place. Thermocouples were placed next to the plastic bottles, so that plastic did not melt onto them and destroy the thermocouple tips. The fires were lit by prescribed burn teams using a drip-torch along the edges of the prescribed fire plots and the fires were allowed to burn towards the regions containing the instrumented cans and bottles. Due to the patchy nature of the low-intensity fire, particularly at Isolation Cottages, not all samples were exposed to burning and some of the cans and bottles remained unmodified. Afterwards, thermocouples and all cans and bottles were photographed in the field, collected, labeled, and brought to the laboratory.

3. Results

3.1. Plastic Bottles

Both laboratory-heated and field-heated plastic bottles deformed in a similar way. The thermocouples placed next to the bottles in the field, which corresponded in form to those generated in the laboratory, gave maximum temperatures that ranged 39–278 °C (Figure 3). The temperatures are largely consistent with the temperature at the surface of the items in the laboratory experiments (Table 1). So although the thermocouples could not be placed directly on the bottles in the field, the temperatures resulting in the field deformations appear relatively consistent with those generated in the lab. Lidded plastic bottles deformed in response to increasing heat flux in a predictable way that we suggest can be arranged into four distinctive stages (Figure 3). Stage 0: a bottle exposed to <0.1 MJ/m² does not have any identifiable deformations. Stage 1a: a bottle exposed to >0.1 MJ/m² reacts by first peeling its label; during stage 1b: when exposed to >0.2 MJ/m² a slight deformation of plastic is present, but the bottle is still airtight. Stage 2: if a lidded bottle experiences > 0.3 MJ/m², then heated air inside the bottle reaches significant pressure that plastic walls that were weakened by heating burst and a small hole forms. Stage 3: for >0.45 MJ/m² the section of the bottle that is the closest to the source of heat melts, forming a large hole, but the other side of the bottle retains its shape (it is still possible to recognise the type of container). Stage 4: for total energies > 1.5 MJ/m² the bottle melts, becoming highly deformed, and no longer retains its shape. If a high enough energy is sustained, then the bottle will eventually disintegrate completely.

We found that the intensity of heating (cumulative heat flux) also plays a role in producing trash deformation; it is especially important for the lowest heat fluxes. For example, little deformation was observed at <10 kW/m² even when exposed for >20 min (sum of 12 MJ/m²), but at >25 kW/m² heat flux the plastic was melted to a liquid in less than 2 min (1.5 MJ/m²).

Bottles heated without a lid deform in similar stages, although the duration or energy required to reach each phase is shifted (

Table 2. Comparison of times required to reach different deformation stages in plastic bottles, with and without a lid, during laboratory experiments.

Sample	Heat Flux [kW/m−2]	Total Heating Time [s]	Lid Present or Absent	Time at which Label Started Peeling [s]	Time at which Plastic Starts Deforming [s]	Time to Lid Popping (Where Present 0 [s])	Time to Melting a Hole [s]	Time to Plastic Clouding [s]
2	10	120	Y	12	35	40	40	53
5	10	120	N	5	7	No Lid	84	62
7	15	30	Y	7	13	19	19	24
13	15	30	N	8	3	No Lid	No Hole	25
9	15	120	Y	7	14	19	19	26
12	15	120	N	5	4	No Lid	23	31
17	20	60	Y	3	10	11	11	18
20	20	60	Y	4	1	16	16	19
21	20	60	N	4	1	No Lid	39	21
22	20	60	N	6	2	No Lid	32	25
24	25	30	Y	3	7	8	8	14
27	25	30	N	2	2	No Lid	26	15
28	25	30	N	2	1	No Lid	28	21

, Figure 4). Bottles without a lid start label peeling (stage 1a) and deforming (1b) more rapidly than lidded bottles; however, they require more energy input to form a large hole (stage 3) or deform entirely (stage 4). Additionally, the bottle cap was observed to require higher heat fluxes/longer time to deform as it is made of a different type of plastic (HDPE as opposed to PET). We observed that the bottle cap would severely deform when heated at 50 kW/m² for 1 min and would completely melt if exposed for 2 min. The bottle caps withstood 25 kW/m² and only deformed slightly; this slight deformation was the result of being exposed to 1,500,000 joules, which greatly exceeded the amount reached in most of these plastic experiments.

3.2. Aluminium Cans

Aluminium can withstand much higher energy levels than plastic, but the cans were also found to deform in a consistent and predictable way during exposure to heating in laboratory and field experiments. Similarly, to the plastic bottles, there appear to be four recognisable stages of deformation in aluminium cans (Figure 5). Stage 0: a can exposed to <30 MJ/m² does not have any identifiable deformations. Stage 1: discoloration of paint is observed (not dirt or condensates from burning material) in cans that experienced 30–60 MJ/m², maximal temperatures not higher than ~200–300 °C, and sustaining temperatures >140 °C for less than a minute. Stage 2: the paint is totally removed in some patches and the bare aluminium exposed in cans experiencing 60–80 MJ/m²; these observations come from cans that recorded maximal temperatures not higher than 500 °C, with a temperature >260 °C sustained for ~30 s. Stage 3: cans show localized deformation of the aluminium, such that it has a bubbly or uneven surface. This surface deformation is linked to energy flux over time of 80–100 MJ/m², but the maximum temperature and duration the cans appeared to be exposed to was not distinct from cans of Stage 2. Stage 4: extensive surface deformation suggestive of aluminium melting: maximal temperatures are >660 °C. Stage 4 cans were only found in the field and we did not observe this in our lab-created heating experiments.

4. Discussion

Our tests suggest that plastic bottles make good markers for the energy flux from wildfires of lower intensities, where they might suggest heating in the range of <25 kW/m² that is only sustained for minutes (Figure 3 and Figure 4). Within this range, plastic Coca-Cola brand bottles visually indicate a four-stage set of changes corresponding to the amount of energy that was locally delivered to them. If more than ~1.5–2 MJ/m² is delivered to the specific location, plastic bottles melt completely, so they are not a practical proxy to monitor the energy of more intense wildfires. It is important to note that plastic deformation depends more strongly on the total energy experienced, not the heat flux/duration/maximal temperature alone (Figure 6). Because of this, the same stage of deformation may be produced by the same heat flux over different durations: e.g., a bottle heated at a heat flux of 10 kW/m² for 0.5 min reaches only stage 1 (peeling of the label), if heated for 1 min it becomes deformed but is still airtight (stage 2) and when heated for two minutes it is no longer airtight (stage 2), but to produce deformation equal to stage 3 requires 20 min of sustained heat flux (Figure 6). The deformation produced and shown in Figure 6 is a result of the cumulative heat flux experienced by a given sample (calculated as heat flux × time, e.g., 20 kW/m² × 1 min = 1.2 MJ/m²); hence, the same heat flux may give different deformation depending on the duration, e.g., 20 kW/m² for 1 min produced stage 3 deformation, but the same heat flux for 20 min was capable of producing stage 4 deformation. Hence, it is important to consider the cumulative heat flux. Additionally, the stage of deformation also depends on whether a bottle was lidded/air-tight before the wildfire, with unlidded bottles being able to survive exposure to slightly higher energy ranges (Figure 4). Whilst we suggest that the general deformation stages are likely to be the same for bottles from different brands, the energy thresholds are likely to be shifted depending on the thickness and to a lesser degree shape of the plastic. This is highlighted by the bottle tops, which are of different plastic to the bottles themselves, and which appeared more thermally resistant than the bottles.

Aluminium cans have the potential to provide a good indication of the energy flux from higher-intensity wildfires as they require significantly more energy to produce any visible deformation; modification of the original form requires >50 kW/m² for an extended amount of time (>20 min). To deform the walls of the can, temperatures need to be elevated above 660 °C. The time experienced at this temperature may be very brief—on the order of couple of seconds based on our measurements (

Table 3. Relationship between deformation stage and duration and the level of heating required to reach those stages of deformation, based on field experiments. IC—Isolation Cottages, FB—Furzebrook Road.

Location	Themo-Couple Number	Decomposition Stage			Max Temp.	>40 °C		>140 °C (Plastic Softening)		>260 °C (Plastic Melting)		>660 °C (Aluminium Melting)
		In Can	Out Can	Near Bottle		Duration	Area Under Curve	Duration	Area Under Curve	Duration	Area Under Curve	Duration	Area Under Curve
		Location of Thermocouple			[°C]	[s]	[s × °C]	[s]	[s × °C]	[s]	[s × °C]	[s]	[s × °C]
Aluminium Cans
IC	13.4	1			89	208	5023	0	0	0	0	0	0
IC	12.2	1			98	274	9631	0	0	0	0	0	0
IC	13.2	1			99	252	7598	0	0	0	0	0	0
IC	13.3	1			131	114	6808	0	0	0	0	0	0
IC	12.4	1			144	276	13,755	10	1430	0	0	0	0
FB	4.3	1			200	641	19,651	38	6593	0	0	0	0
IC	12.3	2			498	204	26,172	67	18,075	30	11,473	0	0
FB	14.2		2		506	179	15,918	40	12,343	24	9290	0	0
FB	14.1	3			335	207	17,133	44	10,306	16	4908	0	0
FB	2.1	3			423	212	21,487	66	18,738	38	13,226	0	0
FB	4.4		4		793	280	54,091	123	47,370	79	38,434	21	15,205
FB	2.2		4		844	252	21,959	36	17,772	27	16,006	12	9257
Plastic Bottles
FB	14.4			2	59	66	1365	0	0	0	0	0	0
FB	14.3			2	80	147	3653	0	0	0	0	0	0
FB	2.4			3	147	159	8521	10	1438	0	0	0	0
FB	4.1			4	495	336	28,719	61	17,386	31	11,943	0	0
FB	4.2			4	618	681	163,614	401	15,009	369	14,403	0	0

). It would be worth additional study to assess how different cans from other brands respond. However, we believe the standard-sized beverage can of those soft drinks made by the Coca-Cola company (i.e., a 330 mL can) are manufactured using the same process, although it may be that paint colour influences the results (e.g., say red for Coca-Cola, green for Sprite, etc.). We hypothesise, however, that our proposed deformation stages and energy thresholds should apply to cans of Coca-Cola company origin.

It is important to note that our measurements were made based on a thermocouple inside the can and therefore the energy flux to the outside of the can was likely higher. Additionally, the fire environment is highly heterogenous; hence, one end or side of an aluminium can may experience different energies to the other. This highlights that trash deformation will be varied across a burned area due to extremely localised conditions in energy regime (cm scale). For example, the same aluminium can might have sections deformed to stages 1 and 3 (Figure 7). Usually, the most affected fragments face up, and the least changed face down. This effect is most noticeable in the case of low-energy fires (or low-energy sections of higher-intensity fires). In higher-energy fires, trash deformation becomes more uniform within and among trash items; e.g., most of the stage 4 cans were rather uniformly damaged. It is important to note that similar patterns are experienced in studies analysing thermocouple data, e.g., [23]. We suggest that where there is enough altered trash across a burned area, it could be used to examine the heterogeneity of imparted energy across the area.

Our experiments show that different trash types cover the range of energies that are expected during a natural fire. Plastic bottles deform in a consistent and observable way even with minimal exposure to heat, aluminium cans cover the middle range, while glass-based trash can be expected to be a good indicator in the most energetic fires. The deformation of glass bottles as a proxy of conditions during wildfire were not tested during this study, because we were unable to reach appropriate temperatures either in the laboratory or during our UK-based field experiments. Additionally, we were unable to locate any heat-deformed glass bottles during our fieldwork during natural wildfires in the south of the UK (e.g., July 2018, Ferndown Common; May 2020, Wareham Forest; April 2025, Canford Heath). However, heat-deformed glass bottles and windows were reported in news/media from higher-intensity wildfires in the USA, Australia and Greece, and therefore glass bottles may also be used to extend this damage severity approach.

In an ideal world, a direct linkage would be able to be made between the physical combustion process and energy release from organic matter during wildfires and their impact on ecosystem elements [1]. Where it would be possible to translate fire intensity directly to impacts on canopy loss, tree mortality and soil effects to name just a few. We suggest that this trash-based approach can provide a cheap and accessible semi-quantitative measure of the energy imparted to ecosystem elements during wildfires in a relatively controlled and reproductible way. This is because certain types of trash are both very common and produced with the same materials all over the world; as such, they can be used as energy indicators between fire types globally. Future work might seek to test a greater range of cans and bottles from different manufacturers. We note that the bottle tops of the most plastic bottles are different plastic to the bottle itself. The bottle tops were more resistant to the heat (taking >25 kW/m² before they began to deform) than the bottles, and therefore it would be worth further assessing bottle tops, as they appeared to survive lower intensities and thus if found deformed might tend to suggest higher energy fluxes. With further development we suggest that observations of the thermal damage/deformation of trash can provide a cheap, rapid, and reliable approach to infer the maximal temperature reached by a material and the total energy flux to specific locations in wildfires. Figure 8 provides a summary thermal proxy guide based on the visual assessment of Coca-Cola brand plastic bottles and aluminium cans that we suggest might aid those seeking to determine the energy imparted to ecosystems at locations during wildfires.

5. Conclusions

We designed a set of laboratory and field experiments to assess whether it might be possible to determine the energy flux from wildfires based on the level of thermal deformation of different types of trash. We selected two materials to test based on field surveys that showed the majority of trash was either aluminium cans or plastic bottles. We used an iCone calorimeter to deliver known heat fluxes over different durations to the trash and also placed the same brand of trash into prescribed fires, where we instrumented them with thermocouples. Our laboratory-deformed trash was largely consistent with that which we deformed in the field during prescribed fires, allowing schematics to be produced that suggest the maximum temperature and the typical cumulative heat flux experienced correspond to different levels of deformation. We suggest that the level of thermal deformation of different types of trash, made of different materials, can be used as a cheap, reasonably reliable and rapid indicator of the likely energy release from wildfires. This trash-based assessment approach can aid us with understanding the heterogeneity of landscape heating by wildfires as part of post-fire surveys aiding with ecological assessments, fire severity surveys and potentially forensic investigations. This could be used to gather further useful numeric data towards comparing the thermal effects between different wildfires both seasonally and within different ecosystems in areas where humans are present and leaving trash.