California Department of Water Resources uses risk simulation model called CalLite to evaluate effects of water management policies.
The California Department of Water Resources (DWR) is using risk simulation to evaluate and demonstrate how different water management policies affect the water supply, environment and other factors.
In 2007 the DWR and U.S. Bureau of Reclamation developed CalLite, a simulation model for the water management of California's Central Valley. It simulates reservoir operations, operations of the California State Water Project and the Central Valley Project, water-delivery decisions, ecosystems and the California Central Valley's hydrology.
In just five minutes, CalLite simulates water conditions in the Central Valley over a span of 82 years. The user can change the program to demonstrate water management actions, like utilizing off-stream storage reservoirs, enacting groundwater management programs and adding new conveyance facilities.
Stakeholders and the public can download CalLite from the DWR's Web site. The front-facing program allows stakeholders to see how decisions affect the Central Valley. Environmentalists, contractors and regulatory boards use the program.
The DWR used GoldSim, a graphical, dynamic simulation program, to develop CalLite. Nazrul Islam, senior engineer of the DWR, said the program's main benefit is that it shows visually how a decision might impact water resources.
"We have a very detailed and complex model that only a small group of people in California can use and interpret the results. The results from that model are very difficult to extract and you need a very skilled and knowledgeable person to do that," Islam said. "On the other hand, CalLite is user-friendly. People can select a couple of options from the graphical user interface of our model, and then run the model and get the results very easily."
Self-education was a driving factor behind CalLite. "We developed this tool for stakeholders to aid them in learning about our system interactions," Islam said. "With this simplified model they can educate themselves and interpret results that we give them from the detailed and complex model."
The DWR creates strategy plans based on the interactive sessions. "CalLite was developed to screen out alternatives easily and quickly. In a particular application, we had many -- approximately 40 -- alternatives from stakeholders and water-resources planners, and from those alternatives we have narrowed it down to six alternatives," Islam said.
According to Rick Kossik, principal of GoldSim Technology Group, the program is a good match for organizations like the DWR because it tracks a complex system that's surrounded by uncertainty. For example, it's uncertain what the rainfall levels in 2010 will be; however, GoldSim takes that into account and tells the DWR what the rainfall level will most likely be. "The outputs that these models produce -- they don't say the answer is seven. They say the answer is somewhere between three and 10 and the most likely value is 7.5," Kossik said.
He said there are three main benefits of GoldSim: It deals with uncertainty; it shows the information in a diagram, so there's a picture of the system that's being modeled; and it can represent changes in time. The pictures of the models are represented as flow charts, which allow users to see how one variable changes another.
"It doesn't force people to look at equations; it says, 'Here's the relationship between parts of a system,'" Kossik said. "In a water-resource system, you would have pictures on the screen that represent different reservoirs and you might have a picture that represents a pipeline and arrows show that water flows from this reservoir, through this pipeline to this reservoir. No equations, you just see at the top level what the system looks like."
Kossik said the technology behind GoldSim is called a Monte Carlo simulation. This type of simulation is named after a casino in Monaco because it's based upon probabilities. A Monte Carlo simulation allows for the model to be run repeatedly and it replaces the uncertain input, like rainfall, with a new number each time in order to determine the likelihood of any particular outcome.
"We run the model 1,000 times, 100 times or 10,000 times and what that produces is possible futures that this system can follow. Because we're uncertain -- we don't know exactly what it's going to do, but we produce a family of predictions of how this system will behave over time and that gives us a full range of possible behavior. And that's a key to the technology," Kossik said.