The World Meteorological Organization (WMO) has just predicted that there is a greater than 66% probability that at least one of the years from 2023 to 2027 will witness a global mean surface temperature rise of 1.5 o C above the average in the years 1850-1900.
Clearly, nobody in the world today lives in the global mean temperature. It then becomes critical to understand what crossing the 1.5 o C threshold crossing will mean for specific regions, including for a country like India.
To do this, we must first understand how these multi-year forecasts are different from the much discussed El Niño forecasts for 2023 as well as from the longer-term climate projections, which the U.N. Intergovernmental Panel on Climate Change (IPCC) released during 2021-22, as its Assessment Report 6.
(In an El Niño year, there is a belt of warm water expanding west to east over the equatorial Pacific Ocean. Complementarily, in a La Niña year, there is a belt of cooler water expanding east to west over the same region.)
What are atmospheric models?
The earth experiences weather, climate, and climate change. As the old saying goes, climate is what one expects and weather is what one gets. The ocean has a high capacity, which means it will need to absorb a large amount of heat for its temperature to change. This does not happen over a few days. So specialists forecast the weather – the fast changes in the order of hours to days – using atmospheric models that work with the sea surface temperature fixed. Weather forecasts also need the atmosphere to be ‘initialised’ to be able to make predictions.
To ‘initialise’ means to have the best estimates of temperature, humidity, winds, and so on at each point of the computer model, using data from weather balloons, satellites, radiosondes, and such. The weather model becomes chaotic after a few days since fast-moving air, with its low heat capacity, tends to shed the memory of the initial conditions.
Climate, on the other hand, is largely driven by the slow changes in the ocean temperature and its heat content. As a result, the climate forecasts, such as for the El Niño and the monsoons, are made with ocean-atmosphere models. The ocean is initialised using data from satellites, ships, moorings in the ocean, and so on. Some climate modes, such as those for the El Niño, can be predicted six to nine months in advance. A land model is also always included at all timescales, going from weather to climate to climate change.
Weather and climate prediction models do not worry about changes in the concentrations of greenhouse gases since they integrate their inputs just for a few days to a few seasons. Beyond a few seasons, we cannot predict the climate with great confidence. It is thus critical to consider the value that multi-year predictions may offer for decision-making for India.
What are ocean-atmosphere models?
On the other hand, the IPCC projections for the forthcoming decades are ocean-atmosphere models, initialised typically starting from the concentrations of greenhouse gases from the pre-industrial era and integrated forward, into the 21st century and beyond. These models need to account for changes in the concentrations of greenhouse gases as well as sunspot changes and volcanic eruptions, along with land use changes. These integrations capture the global mean temperatures very well, since they are mostly driven by greenhouse gases.
Only in 2017, the WMO recognised the need for shorter term predictions to fill a gap between seasonal forecasts and climate projections. A number of modelling centres around the world were entrained to issue decadal forecasts, which are initialised similar to the seasonal climate forecast models; however, they also accommodate the evolution of greenhouse gases and solar radiative forcing (i.e., changes in solar radiation that affect our climate).
Every year since 2020, the WMO has been issuing forecasts for the following year and the average outlook for the following five years. Its ‘Global Annual to Decadal Climate Update’ just released projections for the year 2023 and the five years from 2023 to 2027.
How good are temperature forecasts?
A quick look at the predictions for 2022, made in 2021, and the actual temperatures for 2022 (both shown below) tells the story of the limited skill of predictions beyond a few seasons. The prominent La Niña cooling is severely underestimated in the forecast while the cooling over India has been overestimated. Models are less skilled at forecasting precipitation than at forecasting temperatures. The forecasted rainfall for 2022 has wider error margins than those for the temperatures, including over India.
Global warming is linearly proportional to the amount of carbon dioxide that has been accumulated since the Industrial Revolution. Given the estimated emissions for 2023-2027, the WMO prediction indicates continued cooling over India, with excess rainfall in both summer and winter seasons.
Predictions of surface temperatures for 2022 made in 2021 (above) stand in contrast to the temperatures that actually occurred in 2022 (below). The La Niña cooling in the tropical Pacific Ocean was underestimated in the prediction while the cooling over India was overestimated.
How can India use these multi-year forecasts?
It is clear that the annual to decadal climate updates have some skill over India. It is worth exploring whether downscaling these climate updates over India using regional climate or regional earth-system models will enhance their value for longer-term decisions related to food security, water, energy, healthcare, and other sectors.
Climate projections tend to be affected by model limitations for the coming decade or two. They are also dominated by the uncertainties in the emissions and socioeconomic scenarios that the modellers assume will be the pathway beyond the first few decades.
Squeezing out any value in the WMO predictions to sustainably navigate the resources will be a worthy enterprise for India.
Raghu Murtugudde is a visiting professor at IIT Bombay and an emeritus professor at the University of Maryland.
