Preparing for the worst: how simulation helps communities plan for natural disasters

View of a hurricane off the south of Florida

Natural disasters can be divided into four classes: geophysical, meteorological, hydrological, and climatological, and have the potential to cause loss of life or destruction of the natural environment and private or public infrastructure.

The number of natural disasters has been increasing steadily, but the number of deaths from such disasters hasn’t. In fact, over the last 10 years, the average has been about 45,000 annually. While this is better than before, it is still too many.

While no place is completely safe from natural disasters, some areas are more susceptible than others. Places that are located on the Pacific Ring of Fire, for example, face more earthquakes, volcanic eruptions, and tsunamis than other places. While hurricanes occur over the Atlantic Ocean, Caribbean Sea, and Gulf of Mexico every year. There has also been an increasing trend of wildfires in California and Australia in recent years.

These disasters can reach billions of dollars in damage and reconstruction costs and result in an incomprehensible number of deaths.

To prevent as many deaths as possible and reduce the cost of natural disasters, simulation modeling can play a part. By using a disaster planning simulation model, relief agencies or governments can first try to predict how these events may unfold, then model how best to deal with them and their aftereffects.

This blog post will explore how different types of simulation models, including those for hospital planning, evacuation, and supply chain management, can be used to develop decision-making tools to reduce the impact of natural disasters.

Simulating the impact of natural disasters

It is clear that there are several places where a natural disaster could strike, but what is less clear is when it will happen or how it will unfold. We know, for example, that the world’s largest active volcano will erupt, but exactly when or what the effects will be is unknown.

As the effects of climate change continue to grow, one natural disaster is taking advantage of the dry regions of California and Australia in particular: wildfires. These are becoming more and more frequent, with ever-increasing impacts on people and property. Is it possible to use disaster planning simulation to identify how a wildfire is going to behave depending on where it starts and the wind? Take a look at the model below, which tries to do just this.

Wildfire with wind (source files available)

This is a straightforward model, but one that has potential. It can already be used to start a fire and predict the impact according to the wind direction.

Combining this model with historical data and GIS mapping can help deliver a simulation that can be very useful to a number of different agencies. For example, firefighters can use it to determine where to set up fire breaks, while other emergency agencies can use it to organize evacuations. If this can be done for wildfires, why can’t it be done for volcanic eruptions, hurricanes, tsunamis, and more?

How hospitals can prepare for emergencies using disaster planning simulation

After a disaster, hospitals often become inundated with patients, so it is important to have a system in place that allows the maximum number of patients to be admitted while optimizing resources, including doctors, nurses, and medical supplies.

Simulation is ideal for this type of situation because you can set clearly defined parameters. These can then be adjusted in advance, or even in a real-time situation, to provide the appropriate resources that will ultimately save the most lives. This can be done using AnyLogic as healthcare simulation software, as shown in the developed model below.

Trauma Center (source files available)

But hospitals themselves can face problems during disasters, as during Hurricane Katrina, when one hospital faced a deadly crisis. This is the worst-case scenario, where the hospital that should be a safe place faces its own problems. In this case, the evacuation of a hospital and its transfer to other hospitals also becomes an important part of disaster management. If this can be done effectively, many lives could be saved. Let’s look at evacuation in more detail next.

Simulating evacuations to prepare for natural disasters

In a disaster, the first few minutes are vital for the survival of people. When a disaster strikes, how the evacuation from a building, or even the movement to a safer place within a building, is managed can make the difference between life and death. This is true for many different types of buildings, from single-story to multi-story high-rise.

Below, there is a model of a two-story building with a connecting stairway and the option to disable one emergency exit. Using just this model, you can identify the amount of time needed for an evacuation, find the best route to take, and understand how the number of people in the building will affect the overall time.

Trying to do this in the real world may be expensive, dangerous, or even impossible, whereas evacuation simulation models are low-cost, can be made for a number of different scenarios, and run multiple times.

Evacuation (source files available)

This type of evacuation could not only be modeled for buildings, but also for villages, towns, and even whole cities. In this case, however, we wouldn’t just be looking at pedestrians on foot, but those using cars or public transport. Using AnyLogic’s Road Traffic or Rail Library combined with a GIS map, emergency planners could test different routes with different amounts of traffic at different times.

Dynamic events, such as the destruction of roads, bridges, or even tunnels, could be incorporated into the model. Imagine there was an evacuation plan, combined with an early warning system, that had been modeled for a potential tsunami before 2004 in Indonesia. Could it have prevented all of the death and destruction? Probably not, but could it have saved lives? We can almost categorically answer yes.

How simulation can help prepare supply chains for natural disasters

When people think of disasters, the first thing that comes to their mind isn’t "supply chain," but it should be one of the first things that planners for managing disasters contemplate.

When there is a disaster such as a hurricane, flood, or any other, items such as medicine, food, temporary housing, and other vital products need to be shipped to the disaster area. But what happens if roads are destroyed or flooded? What about if water routes are disrupted by debris?

Perhaps a village or town has been cut off completely from the outside world. It would mean that supplies would not reach the desired place on time.

This is why having a well-built supply chain that has been tested for disruptions and resilience could help mitigate these effects. Often, there are specific supply chains that have been set up by disaster relief agencies in preparation for such events.

These disaster relief supply chains start from the suppliers, whether local or national governments, private citizens, or companies. Then they go to regional distribution warehouses, to local warehouses, and finally to disaster recovery centers. Some supplies may be stored in long-term warehouses, while others might need to be found immediately. So, the twin problems of inventory level and delivery optimization need to be considered when simulating a disaster response supply chain.

Below is a model that simulates product delivery in Europe. The supply chain includes manufacturing facilities and distributors. This is the type of supply chain and logistical model that can be modified to be used in situations of natural disasters.

Product Delivery (source files available)

The future of simulation in natural disaster response

A multi-agent-based approach is a highly effective modeling method for simulating individual interactions in a dynamic system. It stands out for its ability to simulate scenarios with unpredictable behavior. And this is exactly the type of behavior that disasters have.

Developing this type of system in risk-free settings can help with preparing in advance for disasters and then dealing with the event and its aftermath. But creating one simulation for one event by itself may not be enough to understand and deal with all the effects of a disaster.

By developing different disaster planning simulations and integrating them together, disaster management will have as much information available as possible to be prepared to make decisions beforehand, during, and after the disaster event to ultimately save lives and resources.

Don’t forget to subscribe to our monthly newsletter to stay up to date with AnyLogic blog posts and our Model of the Month.

Related posts