Understand complexity to understand weather

Via Scott Thomas (@ScottGL1 on twitter) comes a very interesting note on a US weather service forecast from yesterday:

I live on a small island located in a steep-walled inlet that opens onto an inland sea on the Pacific Coast of North America. Our island is medium-sized, about 12 km long and 8 km wide. Part of it sticks out into the Strait of Georgia, which is part of the larger Salish Sea that exists between Vancouver Island and the mainland. Part of our island sticks into Atl’ka7tsem/Howe Sound, an inlet that leads from the Strait 45kms inland to the mouth of a river valley that drains the Coast Mountains. In winter, katabatic winds can blow more 100km/h onto the north shore of the island, coating it in ice and snow with a -25C windchill while 12kms away on the southern shore of the island, it can be a nice warm, calm, and sunny spring day, where the temperature feels 30 degrees warmer.

If you count on the local weather forecast, which comes from a mere 15 km away at the Vancouver airport, you will have no idea about the weather on Nexwlelexwm/Bowen Island. The Vancouver airport is located on an island in an estuary at the mouth of a long and broad river valley and experiences completely different weather.

It still boggles my mind that people who live where I live fail to grasp the hyperlocal nature of our microclimates. If you rely only on weather apps and have no idea how the forecasts on these apps are made (or indeed what a 60% chance of rain means), then you might think that meteorology is a big lie. In fact, the limited accuracy of long-term weather forecasts is often one of the things that climate change deniers use to bolster their idea that you can’t forecast the weather and you can’t trust the “weather scientists.”

Trying to predict the intensity of an atmospheric river or the landing point of a compact sub-tropical cyclone is an important function for weather forecasters. But it is impossible to tailor forecasts to the hyper local conditions. I know, by virtue of the fact that my house faces southeast, that the gale warnings that come from an atmospheric river forecast are important for me to heed. The rain will fall everywhere, but it will be more intense on the windward-facing slopes and with a 90 or 100 km/h wind gust, it will be driven into the cracks and seams on my house. I have to seal things up if I don;t want leaks. I have to make sure stuff is bolted down or put away and that the fireplace remains dry, as the rain can be driven into my chimney under the cap.

Literally a few hundred meters away, over the ridge behind my house, there will be no wind. Rain, yes. But if you panicked upon hearing the gale warnings, you might be surprised to find that the wind didn’t matter to you at all. People express anger or frustration all the time on our neighbourhood Facebook pages. Sometimes folks will ignore a warning that actually applies to them, because the last one didn’t affect them at all. That lack of situational awareness is perilous and it is not the fault of weather forecasters.

We just do not have a very good sense of how complex systems work or how we are supposed to relate to them. There is a broad societal expectation that experts will give us answers. Weather forecasts do not provide answers, they provide guidance. To use a weather forecast, you have to also participate in sensemaking and decision-making. You have to have situational awareness about where you are and what information you have about your current state, and you have to have an idea about where the forecast information is coming from and what it means. You need to understand the cadence and granularity of the forecast and to know that forecasts about volatile weather systems can change by the minute. With weather emergencies, you need to be able to prepare and take action, even if the outcome isn’t as severe as the forecast made it out to be. And you also have to realize that things could turn out to be worse than the forecast for your area at any given moment.

This weather forecaster, upon retirement, offers us good wisdom for living in a society where we have tools and expertise that help us live with complexity. This little missive reveals what it is we need to do as complexity practitioners and experts in different fields and it also illuminates how to be a better consumer of data about complex situations, whether that is the economy, the weather, our own health or the myriad of other places where the future is just a set of probabilities.