Even as wildfires drain National Guard resources,
the Pentagon is racing to develop computer models
that can better guide decisions about sustainability efforts.

Patrick Tucker / Defense One

 (August 10, 2021) — The Pentagon is in the midst of a massive, multi-year effort to better adapt to climate change and reduce greenhouse emissions. But the changing climate is already imposing costs on the military and even challenging how well it can prepare to fight other nation states. 

“In terms of current operations, we have National Guardsmen, we have active-duty soldiers, we have active-duty airmen right now participating in firefighting support efforts. So these are…folks who are not doing a primary job. So right now we are experiencing climate change and effects. Right now, we know that these are going to only increase over time,” Richard Kidd, the deputy assistant defense secretary for environment and energy resilience, said in an interview. 

That’s just one of the most obvious examples of climate impact on the military. Humanitarian assistance and support for civil authority are also Defense Department missions, as outlined in the National Defense Strategy, and climate change is growing those missions’ size and scope, Kidd said. 

“We have already seen anecdotal evidence for increased demand for domestic support. If you track the number of days the National Guard was needed to provide support for civil authorities, last year was the highest year on record,” he said. 

Part of that was the response to massive protests across the country. But this year, with far fewer protests, “We are on track to exceed that amount. This is National Guardsmen called up to fight forest fires,” he said. “Likewise, the U.S. Army Corps of Engineers response for hurricanes and droughts in support of the national response through FEMA,” will also increase, he said.

Supporting firefighters and other people dealing with natural disasters is satisfying work for a lot of troops, Kidd said. But, he added, “There’s an opportunity cost: if equipment and personnel are being used for that, they aren’t doing other things. They aren’t doing the sort of warfighter training that they need to do.” 

Eventually, dealing with the effects of climate change will become a key area of military involvement, he said. “We absolutely predict that that demand set will only increase, and yes, we can do that.”

More and more security experts agree. 

In June, the International Military Council on Climate and Security released its second report on the impacts of climate change on issues such as governance and civil unrest across the globe. They surveyed experts from a variety of institutions—including the Planetary Security Initiative at the Netherlands Institute of International Relations, the Hague Centre for Strategic Studies, and the French Institute for International and Strategic Affairs—asking them how they expect various risk areas like biodiversity, water availability, and instability within nations to evolve over the next decade. The experts held a dim view. 

“Respondents expect a majority of risks will pose high to catastrophic levels of risk to security. Ten and 20 years from now, respondents expect very high levels of risk along nearly every type of climate security phenomena,” the report said. 

The experts concluded that the global governance system isn’t prepared for many of the risks. So, in part because of that lack of preparedness, more and more of the international response to climate-change-related issues will fall to men and women in uniform.

“Militaries will be increasingly overstretched as climate change intensifies. As the pace and intensity of extreme weather events increases, countries are increasing their reliance on military forces as first responders,” they wrote. “While direct climate change effects regularly threaten military infrastructure and threaten to reduce readiness, the most pressing security threats will come from climate change-induced disruptions to social systems.” 

In 2019, the Pentagon launched a broad review of the effects of climate change on the military. This review put no price tag on current or future costs—in part because they depend on just how much the Defense Department can reduce its own emissions.

“In terms of greenhouse gas emissions, if we were a country, we would be the 55th largest emitter of greenhouse gases by country status,” Kidd said. “The department absolutely recognizes we have to be part of the solution.” 

Reducing emissions at the department’s hundreds of installations, which produce about one-third of its greenhouse gases, may seem to be the easy part, since buildings and infrastructure are simpler to model and modify. But that’s also the problem:  Mistakes are literally cemented. Simple wrong choices can cost millions and the “wrongness” of any choice usually becomes clear only over time.

“Six inches or a foot of a seawall equates to millions of dollars in construction costs,” said Chris Massey, a U.S. Army Corps of Engineers mathematician.

But it’s hard to predict just how much seas will rise and storms will intensify over the next decades. 

Decreasing the U.S. military’s operational carbon footprint is a different sort of game, one of understanding as perfectly as possible how any decision may result in a better or worse outcome, or lives lost. Here, too, a better understanding of the associated risks in trading in a fuel-hungry technology for something greener will be the key. 

This is where new technologies like virtual twinning come into play. High-resolution modeling techniques will make it possible to better understand the effects of climate change and find ways to reduce the department’s carbon footprint. 

Understanding how to better predict the weather has been a Defense Department dream since the end of World War II. But the technologies on hand today have made realizing that dream a real possibility.

Predicting the Weather 

For the U.S. military, the idea of using computers to better model the climate is older than computers themselves. Just after World War II, the U.S. military briefly entertained the idea that it might be possible to precisely predict the weather, and thus control it, through the use of vast electric calculators. As bizarre as the notion sounds, two of the best minds in the country firmly believed it and managed to convince the U.S. Navy to fund their research.

In October 1945, mathematician John Von Neumann—credited with some of the most important computational discoveries in history—and Vladimir Zworykin—the father of the television—marched into the office of Adm. Lewis Strauss and pitched their idea: a machine that could perform calculations quickly enough to account for all the variables in weather and climate, and eventually predict rain or snow as plainly as time. Strauss provided them with $200,000 to construct their machine, which helped pave the way for random access memory, or RAM, and the future of modern computation as we know it. What it did not do is accurately predict the weather. 

While Von Neumann and Zworkin grossly underestimated the number of calculations a machine would need to make in order to perfectly predict weather, they were on to something. More computation allowed for much more precise modeling of weather conditions, which the U.S. military of 2021 is using now to rethink how and where they build installations.

In July, the Army announced that it had entered into a new cooperative research and development agreement with Microsoft to test how the Army’s coastal storm modeling system, CSTORM-M, works in Microsoft’s Azure cloud environment. The hope is that the movement to a massive enterprise-level cloud will allow them to run new, never-before-deployed simulations of the coastal sea rise and also to allow researchers to use the model results to look at coastlines in more detail. 

“To do that we need high-fidelity models,” Massey said. “They are computationally efficient, but at the same time expensive. They require a lot of CPU hours to get that level of accuracy. When it comes to construction cost, accuracy matters.”

C-STORM-M takes about 10,000 historical records of past storms and models them to understand what might happen with future storm sturges and other climate-related weather events. Moving the operation to an enterprise cloud will allow researchers to run the model over and over again much faster, and that should improve the model’s accuracy. It’s the difference between 100 practice hours and 1,000 or more. It will also allow them to bring in other forms of data and further improve understanding of potential probabilities. 

“If for some reason that’s limited, and instead of running a thousand simulations you can only run twenty, well then the uncertain parts in your answers in the whole probability space are much larger. That equates to large dollars when it comes time to actually construct something,” Massey said.

Bruno Sánchez-Andrade Nuño, the principal scientist at Microsoft’s AI for Earth Program, says running the model at higher frequency in the new environment will also improve how well the Army can apply the model in different locations. Incorporating new data layers related to satellite images, infrastructure, etc.. will allow the military to much better understand not only what areas will be underwater or hit by heavy hurricane winds, but also model resiliency. For example, how the shutdown of a particular road might affect traffic, or how long it will take for the base to get back on its feet after a major event.

 In essence, running a high-resolution climate model in an enterprise cloud environment thousands of times is a sort of prediction machine, but instead of perfectly predicting the weather, it offers a sense of what the weather will cost and what problems it will bring, Nuño said.

 “More than a ‘what if’ forecasting tool, [the output] would be an assessment of risks under certain circumstances,” he said. “Any asset, any service, any population is going to be threatened by those storm surges, and so it’s extremely important to understand those risks and then use that,” to inform building decisions. 

The computations, Massey said,  “are giving you the answers to see how high do you need to build? How high does that levy need to be? …Those construction costs, if you are off or you haven’t done enough modeling to really satisfy your accuracy requirements, then that could result in you needing to add six inches or more of uncertainty or engineering safety margins on top of what you’ve already done.”

The Navy is doing similar modeling work for its facilities in Hawaii, as well as its shipyard in  Portsmouth, Virginia, through a process called digital twinning. In essence, they create a very high-fidelity digital mockup of the new base, building, or piece of infrastructure you are trying to build—a virtual twin—then hammer it over and over again with disasters: high winds, too much rain, too little rain. 

“We actually do modeling through digital twins where we can actually look at the probability of an event occurring and the costs associated with mitigating that, and that would actually drive some of the construction work,” said Rear Adm. John Korka, commander of Naval Facilities Engineering Systems Command, during the recent Sea Air Space event in Maryland. 

In April, the Defense Department announced it would make its Climate Assessment Tool, or DCAT, available across the department. The tool allows planners to “look at the effects of climate change on our installations over two time periods, two greenhouse gas emissions scenarios in eight areas, drought, temperature, riverine, sea level rise, these type of activities,” Kidd said.

All of that climate prepping is putting the Defense Department in the lead as a developer and purchaser of microgrid technology, which could help bring down the cost and spur innovation. 

“We want to build installation resilience. How do we do that? Basically on-site microgrid and on-site power generation,” said Kidd. He pointed to reports showing that the Defense Department could make up one-third of the market for microgrid and large-scale power storage due to new resilience requirements stemming from climate change. 

New buildings will also be outfitted with new sensors to feed more data into future modeling, through the smart installations initiative, which should help the Defense Department much better understand how it’s using—and in some cases wasting—power.

The Air Force is also working to reduce emissions through smart infrastructure investments, including building a 240kW solar photovoltaic array at Hurlburt Field in Florida and modernizing 60 buildings (in addition to constructing a solar array) at Moody Air Force Base in Georgia, two projects that the Air Force is undertaking with Schneider Electric.

Reducing Operational Emissions

But installations only account for a third of the Pentagon’s greenhouse gas footprint. The majority comes from operations, where going with greener alternatives can carry. It’s also an area where the Defense Department has led for decades, mostly due to practical concerns about fuel availability. 

The Defense Department will play close attention to how to achieve the same level of capability at lower fuel use, Kidd said. 

“We’re not going to talk about curtailing operations. We’re going to talk about: How do we maximize the impact for the energy that we use, which is in terms of efficiency, but it’s also about artificial intelligence, aircraft planning? So can we be even more precise, even more exact? Can we cover the same set of targets with fewer assets through better use of intelligence and for those assets?” he said.

Here, too, high-fidelity computational models, likely run in large cloud environments, will let the Defense Department find new efficiencies.  And the DOD is already using virtual twinning for new jet designs, such as the experimental sixth-generation fighter jet. 

Other technologies, like solar thin film, have also reached a maturity level that will allow either new, more efficient vehicles or integration into existing ones. 

“There’s a variety of technologies out there, very-thin-film photovoltaics that can be put on the wings of aircraft, principally unmanned aircraft, to give them very extensive range, adding a hybrid drive to many of our vehicles, battery storage, energy capacitors. We have a range of investments and opportunities in the operational energy space that will drive up capability and drive down fuel consumption and greenhouse gas emissions,” Kidd said. 

The Defense Department, he said, is also part of a new Biden administration initiative with the commercial air industry to convert waste streams into new types of fuel. And the Pentagon has long been experimenting with new genetically or molecularly engineered fuel types. 

The goal is to get to a net-zero carbon footprint by the middle of the century. But many things on the battlefield still can’t be easily converted. “Add all that up…and we’ll probably still be burning liquid fuel in 2050,” Kidd said. 

That means that the Pentagon will have to find some way to offset the fuel, which Kidd said could mean using Defense Department land differently. 

“We have a tremendous amount of land. We can change the way we use that land, trap more carbon in the soils or vegetation, to serve as a sequester offset for the fuels we are still burning,” he said.

Still, there’s no way to model for absolutely every eventuality, especially in the context of possible conflict. That means military leaders will continue to be cautious in terms of technologies they try out or deploy, especially if it comes at the expense of a capability they know works. So convincing military leaders that new solutions work better than what exists today will be difficult.

“I feel comfortable that we can adapt, but I don’t know the magnitude we can adapt to,” Korka said. “We talk about the modeling, again, you’re making a decision based on the probability of occurrence, and that’s what you’re putting in. But what if you get it wrong? And what if you get it wrong with something that’s mission-critical?”

Posted in accordance with Title 17, Section 107, US Code, for noncommercial, educational purposes.