Simulating REalistic HIgh-resolution PREcipitation in the tropics

Periodic Reporting for period 1 - REHIPRE (Simulating REalistic HIgh-resolution PREcipitation in the tropics)

Berichtszeitraum: 2017-10-01 bis 2019-09-30

The Maritime Continent an intricate tropical archipelago with very complex topography and surrounded by one of the warmest oceans in the world (Indo-Pacific warm pool). This combination of factors creates very specific precipitation features and makes the region extremely problematic in terms of simulating the physical mechanisms that generate rainfall. In fact, a realistic representation of deep convection and rainfall in tropical regions remains a major challenge for atmospheric models and they often incur in systematic errors. The Maritime Continent is at a centre of interaction across scales, which connects local processes with the global circulation. Hence, global model errors in this region propagate through the entire Earth system and affect model estimates in remote regions too.

This project addresses the question of whether precipitation is more realistically simulated by providing finer description of the atmosphere, representing convection explicitly and taking into account air-sea mesoscale interactions.

A better understanding of convective processes and their interaction with the environment is crucial to enhance the realism of tropical rainfall in models and improve our estimates of changes in rainfall characteristics in the Maritime Continent. This is key to determine effects on agriculture, forest fires, floods and vector-borne diseases in a region that is home of 275 million people and with climate repercussions at global scales.

The overarching aim of the project is advance our knowledge of the interactions between local circulation patterns, fine-scale air-sea interaction and tropical convection to represent precipitation characteristics more accurately. This is achieved through three specific objectives: 1) To determine the effect of model resolution on local circulation, convective processes and precipitation, 2) To identify key mechanisms required to correctly represent local circulation and convection and 3) To establish the implications of future climate change for convective processes and rainfall.

This project has completed its objective through a combination of next-generation modelling tools, advanced statistical techniques and a breadth of observations. During the project lifetime, a convection-permitting ocean-atmosphere coupled system was configured for the region of interest and a range of model experiments were completed at one of the most powerful supercomputers. Sophisticated technical and statistical approaches were used to distill useful information from the massive amount of data generated by the modelling system. A range of observational products including satellite-derived estimates, gauge measurements and radar information were used to provide a backdrop for model comparison and evaluate its performance.
Finally, information from multiple Global Climate Models was gathered to generate a climate change signal used to force the fine-scale modelling system and establish the future evolution of rainfall and related processes.

As a result of this, we have shown the adequacy of the atmospheric modelling system to study rainfall and convection in the Maritime Continent. We established the need of very high-resolution modelling techniques and representing convection explicitly to realistically simulate precipitation in the tropics. We showed that the model is able to capture the main characteristics of rainfall in the region, both over land and over the ocean. Challenges remain in terms of the excess of rainfall produced by the model over land, an issue shared by other modelling systems, while the key aspects such as the diurnal cycle were clearly improved at convection-permitting scales.

We identified mechanisms that contribute to correctly represent rainfall-generating processes in the region such as the moisture convergence on land due to water vapor availability in the lower levels of the atmosphere, the vertical structure of deep clouds and the generation of non-precipitating shallow convection. We show the limited effect of resolution and convection on the representation of sea breeze and thus concluded that differences in rainfall estimates are not produced by how the model represents this local circulation.

We found that the use of ocean-atmosphere regional climate models in the region requires further development since the sea surface temperature was prone to systematic biases that in turn lead to substantial precipitation errors. However, we also showed that the diurnal variation of sea surface temperature may play a non-negligible role in modulating rainfall processes and thus should be taken into account when studying convection in the region.

We also explored the relationship between temperature, dew-point temperature and precipitation extremes in the Maritime Continent using a convection-permitting model and multi-model climate change signal to quantify its future evolution. We concluded that near-surface temperature or dew-point temperature, and their changes, are not representative of the entire column temperature and its changes, which may produce apparent inconsistencies with the Clausius-Clapeyron scaling for precipitation.

During this project, I published three peer-reviewed articles in international journals and submitted other three that are currently under review. I also made contributions to 3 international conferences, 1 international workshop and 2 seminars, and was invited to two talks aimed at the general public. Overall, the project produced enormous amount of data (140TB) and code, some of which were shared through public repositories. Videos and written pieces explaining aspects of deep convection in the Maritime Continent were shared through the project website and my personal website. I have visited three world renowned institutions in the US (NCAR), France (CNRS) and Australia (UNSW) and collaborated with multiple international researchers these and other institutions (Newcastle University, KNMI).

For the first time, this project has set up a convection-permitting atmosphere-ocean coupled modelling system to investigate the key factors in simulating realistic convective processes and precipitation in the Maritime Continent. This technique allowed us to quantify the importance of explicitly resolving convection in the representation of rainfall characteristics in the region. It also allowed us to establish the role of convective representation on aspects such as the vertical structure of the atmosphere, the moisture availability in the lower levels and the formation of non-precipitating shallow convection, which were found to be crucial in the correct simulation of rainfall processes. Other aspects that were originally deemed also critical, such as the representation of sea-breeze, were not captured differently by our model experiments and thus had only limited impact on the improved realism of precipitation. We also explored the future of precipitation extremes in the Maritime Continent and how different the estimates may be in our modelling system compared to standard model configurations. This was done in the framework of precipitation scaling, where extreme rainfall changes are related to projected warming. We found that the intensity of short-timed events is likely to significantly increase in most land areas of the Maritime Continent.

Convective clouds and rainfall represented by the model

Time of maximum precipitation in observations and model experiments

Cross sections of temperature, humidity, winds and cloud mixing ratio in model simulations

Periodic Reporting for period 1 - REHIPRE (Simulating REalistic HIgh-resolution PREcipitation in the tropics)

Diese Seite teilen

Herunterladen