Do we need climate models at all?

At a student’s presentation recently, a climate faculty again expressed total distrust of models. This is surprisingly common. Researchers who collect and work with data think models are not trustworthy. The only problem is that the so-called ‘data’ is often generated with help of a thousand ‘models.’

All satellite data is produced by converting the signal received by the satellite into data by removing the noise and other interferences using models; commonly referred to as correction algorithms. And no real-life decision is made without a model — a mental model. Whether it is buying a cake or a car, we use mental models based on prior subjective and objective information to decide which place has the best cakes and which company makes the best or most reliable cars.

So why the cynicism about models? May be understanding what models actually do can help. The earth’s climate system is driven nearly completely by the energy coming from the sun. Of course there are internal sources like volcanoes or the greenhouse gases being produced by human activities. But these internal sources also affect climate via the so-called radiative forcing or the energy balance of the earth’s climate system.

What is really amazing about climate modes is that they can take this energy coming in at the top of the atmosphere and produce nearly all features of earth’s climate – temperature, humidity, winds, rain, sleet, snow, monsoons, El Niños, ocean currents, river flows, and what have you. And now we are even getting really good at depicting life from phytoplankton and zooplankton to fish as well different types of vegetation and biodiversity. So why the cynicism? Because models tend to be imperfect.

But if you focus on the imperfections you will miss the woods for the trees. It’s like saying you will abandon the car because it is pulling to one side or is producing some funny sound. Just as you would drive your car carefully to the mechanic to get it fixed, climate models are constantly pulling to one side or producing funny features. But they are also being improved constantly. The progress we have made in each decade since the 1960s, when large-scale modelling began, is quantifiable. The stunning skill level of weather forecasts is clear evidence.

Climate models are driven by the same underlying physical principles as weather. But while weather forecasts can be made several times a day to learn the shortcomings of the models and to improve them, we can make monsoon forecasts only once a year while El Niño forecasts can be tested only once in 2-7 years. And thus improving climate models takes time. No modeller however believes that it can be done without sustained high-quality data.

Data alone won’t help

Can we understand all the weather and climate problems by just using data? Not possible at all because we simply don’t have enough data. We never will. Even if some particular problem can be resolved by data, we need models to say what will happen in the future — either at weather timescale or at climate timescale.

More importantly, the methodologies to merge data and models are now yielding unprecedented insights into our climate systems. Neither the data nor the models can complete the understanding by themselves. Models informed by, tested against and validated with data, are our navigation tools for tomorrow, next year and the next decade.

It’s quite common for climate scientists who work exclusively with data to say that a modeller just cannot be complete without working with data. This is true but a modeller cannot really get too far without testing the models against data. But it may be important for data scientists to work with models to get over the fear of models.

It is also clear now to the adventurous data experts that models are the best tools to plan data-gathering efforts in terms of strategising the spatial and temporal scales at which data needs to be collected to get the best bang for the buck for each bit of data collected. Especially when you consider the cost of a satellite or a ship used for observational efforts.

Data are thus invaluable for making progress on understanding many of the weather and climate processes but models are indispensable as well. Just as many cars get recalled for their imperfections but get fixed and put back on the road, climate models will keep improving.

The writer is Professor, Dept of Atmospheric and Oceanic Science, University of Maryland, US

COMMENT NOW