On this Page
The CLIVAR Climate Dynamics Panel 5th Annual Workshop was organized at the Mantra Lorne near Melbourne, Australia, during 24-27 February 2025. The workshop brought together researchers from across the globe to dwell on the impact of small-scale processes on weather and climate. The focus was on small-scale atmospheric, oceanic, and land processes. The effects of climate change on weather and weather systems were also discussed, and deliberations were carried out on bridging the gap between weather and climate. This post will dwell on some aspects of the workshop that the CMIP community can utilize to bridge the gap between weather and climate. I will use the examples of the small-scale processes that were discussed in the Annual Workshop. Many thanks to all the workshop participants and the fruitful discussions, which have helped me put together this post, the CMIP-IPO, and the ICPO, whose support enabled my attendance at the meeting.
Importance of small-scale weather processes
The impact of anthropogenic activities on global-mean climate change is unequivocal. However, climate projections from models participating in the Coupled Model Intercomparison Project (CMIP) remain uncertain on regional scales. This is perhaps due to the contribution of small-scale weather processes (SWP), which coarse-resolution climate models do not resolve satisfactorily. The contribution of such weather processes to internal variability in the climate system increases from global to regional scales. This larger internal variability at regional scales is a significant source of uncertainty in model projections. This uncertainty in regional projections is unlikely to change in the medium term. Despite many advancements in CMIP6 model physics and model resolution, even finer resolution and better parameterization might be essential to capture the small-scale processes of the weather system.
Better parameterizations using km-scale simulations
The climate over a region can be considered an aggregate of the various weather systems in that region. For instance, the summer monsoon rainfall over India (the country I come from) is the sum total of rainfall caused by various SWPs during the rainy season. These systems can have varied spatial and temporal scales, ranging from short-lived localized convection to synoptic-scale low-pressure systems and intra-seasonal oscillations. Therefore, if we aim to get reliable regional projections, we must strive to improve the representation of these SWPs. This could be done by increasing the model resolution. Computational constraints pose a limitation on how fine we can go. Regional downscaling is another approach. However, increasing the model resolution alone is not sufficient. SWPs interact with large-scale climate drivers as well. Identifying the causal pathways of such interactions between weather and climate systems and assessing them in climate models is vital to get the representation right. To account for sub-grid scale processes, climate models parameterize many of them – for example, deep convection parameterization is used to approximate tall cumulonimbus clouds. These parameterizations can lead to significant biases in rainfall. Clouds also play an important role in altering tropical circulation, and are reported to affect large-scale phenomena such as ENSO. Atmospheric model parameterizations do not capture such feedback very well. Specialized models that can produce kilometre-scale simulations, without using convection parameterization, are valuable tools to evaluate causal pathways between weather and climate and understand the internal mechanisms of such processes.
Caution while interpreting regional impacts
Low-pressure systems are an important source of rainfall across the globe and are often associated with extreme precipitation events. Identifying long-term trends in these systems can help identify vulnerabilities. Coupled models have a significantly biased representation of such systems, which can impact the simulated trends in regional rainfall. The understanding of such systems can be enhanced by dynamical downscaling to some extent. Variations in tropical moist margins impact weather systems such as lows at synoptic time scales, and their movement is controlled by horizontal advection.
Land processes and land-atmosphere interactions impact predictability at sub-seasonal to seasonal (S2S) time scales due to the long-term memory of soil moisture, snow cover, etc. Coupled models show unrealistically strong coupling between soil moisture and precipitation. Large-scale deforestation can impact evapotranspiration, moisture convergence, and large-scale circulation, ultimately impacting precipitation. Studies show that the impact of land-surface processes on precipitation is different in storm-resolving models compared to coarse-resolution coupled models. Clear-sky conditions are beneficial for plants, enhancing the gross primary productivity. Frequent and long-lived storms can reduce solar insolation, and hence, they need to be resolved better in climate models. All these aspects stress that we must be cautious when interpreting regional impacts.
Examples of SWP interaction with seasonal mean
In mid-latitudes, surface weather is tightly coupled to the propagation of atmospheric waves aloft. The simulations of atmospheric waviness improve with high-resolution models, such as those participating in the HighResMIP. Oceanic processes, such as the western boundary currents, are also reported to impact atmospheric waviness. In a recent study, we found a strong association between mesoscale oceanic eddies and monsoon rainfall in India. These eddies are often referred to as the “weather” of the oceans, and the associated SST perturbations modify the surface wind. Such mesoscale air-sea interactions dominate in many regions across the globe. Representing small-scale systems in the oceans, such as eddies and fronts, is quite important from a regional perspective. I presented an example of the land-ocean-atmosphere interaction via terrestrial rivers. Climate projections have considered the role of freshwater forcing from rivers in modulating the climate system. However, their impact on S2S scale variability is not quantified. I coupled a river-routing model to a GCM to determine whether the Indian Monsoon simulation improves. As it turns out, when you have interactive freshwater fluxes from the rivers in a GCM, the S2S variability, and the seasonal forecast skill of the model at simulating the seasonal mean monsoon rainfall improve.
Small-scale processes can be considered the building blocks of the climate systems. They can provide feedback to the large-scale via scale interactions. In the first iteration of my analysis of the coupled model with river routing, I found a significant improvement in the simulation of tropical lows. Subsequently, I found that these synoptic-scale lows feed back to the intra-seasonal time scales by impacting tropical intra-seasonal oscillations, impacting the seasonal mean. If I had extended this analysis, the scale interactions might have impacted even longer time scales. This is a classic example of scale interactions, something to watch out for in our endeavour to bridge the gap between weather and climate!
Taking action to improve the representation of small-scale processes with big prediction impacts
The small-scale processes of the Earth system can impact predictability at many time-scales – from weather to sub-seasonal to decadal forecasts and climate projections. This means there is a need to build collaborative activities across the broad framework of observations, experimentation, predictions, and projections. The World Climate Research Programme is well-placed to address some of these challenges by deriving synergies between its core projects and lighthouse activities on small-scale processes.
To summarize, the relatively “coarse” resolution of GCMs might seem to be a roadblock in bridging the gap between weather and climate. We also need to be cautious when assessing the regional impacts of climate change. However, some of these challenges can be overcome by identifying the causal pathways between weather and climate. This could be achieved with our understanding from km-scale simulations, better parameterizations, breaking down the climate into weather systems, and identifying and understanding the scale interactions.
My suggestions for the CMIP community action for improving the representation of small-scale weather processes in climate models are:
- Develop better parameterizations by utilizing the knowledge from km-scale simulations
- Exercise caution while deriving conclusions at regional scales, as SWPs influence regional-scale climate strongly.
- Rigorous assessment of the scale-interactions in CMIP simulations to quantify regional uncertainty.
