Interview with an Expert: Marc Castellnou

At the height of the pandemic, the Robotto team sat down with wildfire expert Marc Castellnou to learn more about issues and challenges firefighters face when dealing with wildfires.

Marc Castellnou is President Patronat at the Pau Costa Foundation, Inspector Cap de l’Area Forestal with Bombers Generalitat, Associate Professor at the University of Lleida, and has published a number of studies focusing on the behavior of wildfires. With this expertise, Mr. Castellnou took our team through, what he has identified, to be the biggest challenges facing wildfire analytics and firefighting teams today. His input has driven our team to create software that specifically targets these pain points, helping his team in Spain focus on what’s important- keeping their community safe.

THE VALUE OF INFORMATION

Mr. Castellnou began by stating that there is currently an overflow of information. Firefighters are struggling to understand wildfire behavior by using all methods of data collection and analysis available to them. Today’s method includes sifting through information via radio, aerial monitoring, and personal point of view reports. This process creates a messy and thick data flow making it difficult to have a quality overview of the situation. In addition to this, community concern and rampant misinformation on social media has begun to pull their attention away from tackling the flames, as this parallel emergency requires immediate attention as to not cause panic within the community.

HOW DATA COLLECTION WORKS NOW

Visualization of mathematicall analysis provided by Marc Castellnou

The team was particularly interested in learning how fire departments currently analyze wildfires, and what information sources are consulted during this process. Mr. Catellnou outlined a five-step process in which various aspects of the fire service are utilized in an attempt to gain a better overview.

1. Aerial teams ascertain an overview from above (fire start + 21 minutes)

2. The on-site crew reports the fire’s behavior to command (fire start + 34 minutes)

3. Fire analysts attempt to create a clear view of the scenario using mathematical polygons of fire potential, trying to understand what fire wants to do

4. With this information department leaders identify where efforts will be made and where the fire naturally will spread due to topology and fuel sources

5. Plans are then communicated to the community alongside fire updates and changes in fire suppression methods due to unpredicted fire behavior

FIRE BEHAVIOR PREDICTIONS

Mr. Castellnou spoke about a fire in which teams predicted westward movement as that was the indicated wind direction. This, however, did not happen as the fire was located in a canyon, creating an unpredictable situation where the fire moved north, surprising fire services. It is specifically here AI, according to Castellnou, has the potential to lend a hand, qualifying information and checking assumptions to give fire services a better understanding of what they are dealing with.

Graph of Connectivity provided by Marc Castellnou

Another method of influencing fire behavior currently used by the fire service is the graph of connectivity. Using this teams decide how to prioritize efforts to minimize burn sizes. The effectiveness of this, however, is dependent on the quality of aerial information coupled with quality mapping (including elevation, topology, and fuel types).

COMMUNICATING WITH THE COMMUNITY

Fire services are servants of their communities, and keeping their communities informed in times of disaster is a fundamental aspect of their job. With the growing number and intensity of wildfires, trust in their abilities has begun to wane, making community outreach and increasing credibility vital.

New methods of communication must remove uncertainty and increase the credibility within the local community of the fire service. Castellnou explained that “if you can not tell what you are doing and how you are doing it, you won’t get credibility”. This concern is particularly important in the age of broadcast and social media, as communities have begun to question the credibility and ability of fire services to do their jobs, increasing their active concerns which in turn detract from the focus and use precious man-hours as they work to contain the spread. However, Castellnou notes that if you address the community with an active fire stating “we are doing all that we can” the lack of details does nothing to calm fears.

To present the community with information that will achieve this goal (specifics of what the team is doing, where they see the fire status in the morning etc.) requires advanced modeling and data surveillance methods.

THE PROBLEM

“Within emergency services there are two problems, the one we as service members know we have and the one our bosses, our politicians, think we have” Mr. Castellnou stated, continuing to note that many believe their problem is increasing external credibility. This is of course important, but to do this, service members on the ground need improved fire data that allows them to get ahead of the problem.

What is the solution to this? Many in the fire service (who are not on the ground) believe the solution is more recourses, but according to Mr. Castellnou, this isn’t completely true. The solution according to this GRAF team lead is a better understanding, better information that allows them to create a successful strategy, a strategy that isn’t based on following the fire.

WHAT NEW TECHNOLOGY NEEDS TO PROVIDE

Our goal at Robotto is to create software that equips firefighters with the quality information they deserve. We do this by simplifying the fire surveillance and analysis process, allowing them to gain a better understanding of the problem at hand and keep their communities safe and satisfied with their abilities. As this is our main purpose, the team posed Mr. Castellnou a series of questions, during which he outlined the main goal any new software must achieve: Solutions need to explain the information (not raw data). The software should translate the information is observed into usable, actionable information that allows firefighters to best apply their know-how and resources.

In the year since this workshop took place, we have worked hard to ensure that our platform does exactly this: provide analyzed, actionable data that gives firefighters the information they need to apply their resources and know-how effectively therein by protecting their communities. This, in turn, keeps their communities calm as teams are able to not only get the job done but also communicate effectively, informing their communities on what they know about the fire and what they plan to do in order to keep them safe.

Two new studies highlight need for innovative firefighting technologies

Two new studies published this week outline how the wildfires exacerbate climate change and how in turn, climate change exacerbates wildfires. It’s a vicious cycle that continues to harm not only nature but society as a whole. In the following post, we outline what we at Robotto noted in both of these studies, as well as where we see Robotto Fire’s part in finding a solution.

This week a new study was published on nature.com looking at changes in night-time fires, and the overall impact on the growing number and intensity of wildfires. In addition to this, the UN Environment Programme released a rapid response assessment, outlining extreme wildfire changes and predictions. At Robotto using technology to enhance firefighting, equipping firefighters with vital information in real-time is at the heart of what we do. Applying advanced AI systems to this global existential crisis stems from an understanding of how the vicious cycle wildfires have on climate change demands change. Reports such as the night-time fire study and the UN Environment Programme’s report only strengthen our commitment to using AI for good.

In this article, we outline how each of these reports notes not only the urgency of applying advanced technology to wildfire surveillance but also the importance of changing our reliance on current methods.

Only 1o.4% of wildfire detection occurred during night-time hours between the years of 2003 and 2020.

NIGHT-TIME SUPPRESSION

Night-time suppression is widely accepted as the most advantageous time to suppress flames as this is when wildfires tend to “lie down and be less intense”. (Balch et al., 2020.) While this is widely known and accepted, only 1o.4% of wildfire detection occurred during night-time hours between the years of 2003 and 2020.

WHAT’S CHANGED?

Using the vapor pressure deficit (VDP), or rather, the absolute measure between the air’s water vapor content and its saturation value, researchers could understand the earth’s own ability to aid in containing the spread of wildfires, as higher levels of vapor content aid in combatting wildfire. Using VDP it was ascertained that “47% of burnable lands had a significant positive trend in the number of night-time hours above the vapor pressure deficit threshold, an increase of 80% in total" (Balch et al., 2020.,. p. 4)

In addition to this positive trend, the study noted that "annual mean VDPmin increased by an average of 25% across the global land surface over the 42- year period." This was an increase of 82% in 59% of burnable areas. Fires grow faster between midday to afternoon, when weather conditions proxy maximum fire spread rates and burn intensity.

The number of nights above the vapor pressure deficit has also increased rapidly, most notably "in the western United States (where) the number of flammable nights has increased by 11 - representing a 45% increase over the four decades" Such an increase may seem small in numbers, but on the ground, this has allowed for wildfire season to grow from a summer phenomenon to one that affects us even in winter months. (This winter’s wildfire in Colorado for example.)

"Such an increase in the length of the temporal window that lacks a night-time fire barrier represents an important and overlooked driver that can enable longer duration and consequently larger fire events" (Balch et al., 2020., p. 4.)

WHY IS THIS IMPORTANT?

Increased fuel and thus momentum during the day, and the rapid increase in nighttime VDP, underscores why it is vital we take advantage of nighttime firefighting. To do so we must move beyond limiting fire surveillance methods such as helicopters and planes and utilize the technology available to us, ensuring accuracy and effectiveness during these precious nighttime hours.

If you’re interested in reading more about this specific study, you can find a synopsis by the Washington Post here

RAPID RESPONSE ASSESSMENT

On the heels of the night-time fire study, comes the United Nations Environment Programme's assessment outlining the disastrous effect wildfires have had on our environment: “Spreading like wildfire: The rising threat of extraordinary landscape fires”.

The report, written by 50 experts from private industries, governments, and research facilities, outlines the vast impact wildfires have on not only the ecosystem but our society as a whole. In addition to this the study, using predictive modeling, outlines how wildfires will continue to grow without proper investment in both preventing extraordinary wildfires, as well as containing fires at an earlier stage.

Specifically, the report finds that wildfires will continue to grow not only in size but in frequency.

THE CLIMATE

Wildfires are not only affected by the warming of our climate but also exacerbate the situation. In areas such as peatland and rainforests, which “store large amounts of incorrigible terrestrial carbon” wildfires release vast quantities of CO2, accelerating climate change. This then contributes to the “positive feedback loop in the carbon cycle, making it more difficult to halt rising temperatures”. (United Nations Environment Programme, GRID Arendal a UNDP Program, 2022, p. 10) This cycle becomes a cycle where the wildfires become more intense due to climate change, meanwhile also contributing vast amounts of emissions worsening climate change.

The report states that “over the last decade, it appears that more wildfires are occurring, not only in regions where seasonal fires are common but also in areas where fires do not normally occur. For example, eastern Australia and the west coast of the United States of America generally experience frequent summer fires, but in the 2019-2020 fire season saw record-breaking numbers and extent of wildfires in the regions. The Arctic and the Amazon, however – areas not generally prone to extensive wildfires – experienced record-breaking blazes in recent years.” (United Nations Environment Programme, GRID Arendal a UNDP Program, 2022, p. 26)

Changes to the global temperature and thus the climate has also altered how fire behaves. Fire variables such as weather conditions, fire regimes, land management practices, and ecophysiology of biomass fuel species change to adapt to warmer climates. (United Nations Environment Programme, GRID Arendal a UNDP Program, 2022, p. 34) This disastrous circular relationship between increased wildfires and intensity, and climate change puts our earth and thus our future at great risk.

KEY TAKEAWAYS

The report outlines that current technology has limitations in that they are highly dependent on prevailing weather, fuel conditions, and accessibility. On this, we agree, as current methods of fire surveillance are dependent on helicopters and planes as well as human interpretation. These methods are not only unable to operate at night, come at a high cost, and are prone to inaccuracies.

In the past Robotto Fire has been accessed to be greenwashing. This is an accusation that we take very seriously, as using technology to help solve the climate crisis is at the very core of who we are at Robotto. How wildfires contribute to the continuous warming of our planet is an existential crisis to which we must find a solution. We believe full-heartedly that drones, specifically AI-powered drone software, will be a part of the solution.

One might ask, how does the technology contribute? As the technology empowers fire departments to survey wildfires faster, including at night and in hard-to-reach areas. Having this information during the nighttime hours makes containment efforts more effective- taking firefighters out of the defensive, allowing them to get the job done faster.

The increased effectiveness of firefighters directly impacts the technology’s environmental impact. Through case studies, interviews, and predictive modeling the following impact has been identified:

When Robotto Fire is applied during the early stages of a wildfire, the increase in data (location, size, direction, etc.) enables firefighters to contain the spread faster. When compared to historical fire data, models show an average decrease in burn sizes by approximately 60%.

Containing the fire faster also means fewer man-hours. In our calculations, we predict 50-70% fewer man-hours than currently used.

Finally, limiting the fire results in smaller burn sizes. This of course has an impact on a fire’s CO2 emissions. Models here show a reduction of 60%.

REFERENCES

Balch, J.K., Abatzoglou, J.T., Joseph, M.B. et al. Warming weakens the night-time barrier to global fire. Nature 602, 442–448 (2022). https://doi.org/10.1038/s41586-021-04325-1

Spreading like Wildfire. A Rapid Response Assessment. (United Nations Environment Programme, GRID Arendal a UNDP Program, 2022)

Co-Founder and CEO sits down with CleanTechWatch

This past week our co-founder and CEO Kenneth Richard Giepel sat down with CleanTechWatch, a danish online media source dedicated to identifying how technology can help improve sustainability.

In the article, Kenneth pointed to the immense potential impact Robotto Fire has, reiterating the results from our study where it was found that the technology has the ability to reduce burn sizes by 60% and save 50-70% of man-hours. In terms of sustainability, Giepel noted that the potentially saved emissions equal the same amount as if Denmark were to eliminate all emissions for a year.

The article also touches upon the commercialization of the technology, noting that Robotto is a first mover in the market, putting the software lightyears ahead of potential competitors in the field.

Here is a short paragraph from CleantechWatch’s article written by Josephine Buur Bach translated:

Going forward, it's potentially more than just firefighting, that Kenneth Richard Geipel and his colleagues can contribute to making the world a bit greener.

“At the end of March, we will launch two new use-cases, Search and Rescue as well as Disaster Mapping, both of which operate on the same drone. After which we will focus on the maritime industry, where there is a big problem with oil spills as well as detection of fuel from ships. Here our customers will be ports, who are interested in protecting their ecology.”

Read the entire article for free with a 14-day trial here (in danish)