Image: © Vira Pogromska/Stock.adobe.com
Dr Basanta Kumar Samala tells us about his weather forecasting research, which he says will have implications on renewable energy applications.
Ireland’s efforts to reduce reliance on fossil fuels continues to be a struggle. Despite some achievements in the pursuit of a primarily renewable supply of energy, Ireland is still heavily dependent on fossil fuels.
A report from the Sustainable Energy Authority of Ireland (SEAI) in May estimated that Ireland’s energy-related emissions reached their lowest point in 30 years, but also noted that more than 80pc of the country’s energy still comes from fossil fuels. A more recent report from the organisation warned that despite some improvements, trends in areas such as transport and home-heating emissions could reverse the country’s climate progress.
But with a majority of renewable energy being weather-dependent (ie wind and solar), researchers have been working on ways to improve forecasting capabilities in order to benefit effective renewable energy usage.
One such researcher is Dr Basanta Kumar Samala, a computational scientist who has worked at the Irish Centre for High-End Computing (ICHEC) at the University of Galway since 2016. Over the past few years, Samala has been part of a project focused on creating an integrated solar, wind and wave energy forecasting system for Ireland.
With a PhD in atmospheric sciences and extensive experience in weather forecasting and ocean state forecasting, Samala was well-equipped to take on the project, which has been supported by SEAI, the Department of Housing, Planning and Local Government (Met Éireann) and The Marine Institute under the SEAI Research, Development and Demonstration Funding Programme 2018.
Dr Basanta Kumar Samala. Image: SEAI
Stronger together
“Seamless weather prediction is both an opportunity and a challenge facing the global meteorological community,” Samala tells SiliconRepublic.com.
As Samala explains, the current weather forecast model used by Met Éireann is the Harmonie-Arome model, which provides regional atmospheric data such as temperature, humidity and wind direction and speed.
At the Marine Institute, the Simulating Waves Nearshore (SWAN) model is used for forecasting wave parameters. To calculate wave boundaries, this model uses the WaveWatch 3 framework.
“In addition to wave forecasts, the Marine Institute uses the Regional Ocean Modelling System (ROMS) to forecast ocean state parameters such as ocean temperature, salinity, sea level and ocean currents,” says Samala. “These models (ocean, atmosphere and wave) are run independently each day with different initial and boundary conditions. These varied conditions lead to discrepancies in the forecasts.”
When forecasting agencies compile information for a global weather forecast, Samala says it is common for them to use a combination of climate models referred to as coupled global circulation models (CGCMs).
“Like CGCMs, regional coupled (atmosphere-ocean-wave) models can enable more accurate regional weather forecasts,” he explains.
This is where Samala’s project comes in. Under the supervision of the ICHEC’s Dr Paul Nolan, his research aims to develop a coupled forecasting model that uses Harmonie-Arome (atmosphere), WaveWatch 3 (wave) and ROMS (ocean) in Ireland. To do this, Samala is utilising the Oasis3-MCT coupler, which is a piece of software that allows synchronised exchanges of coupling information between numerical codes representing different components of the climate system.
“Regional coupled models have been proven to be more accurate than their standalone counterparts,” says Samala. “This is especially true for major or anomalous weather events such as storms and other high-impact events, which have been shown to be better predicted by coupled models in terms of both intensity and direction.”
But if coupled forecasting models are generally more accurate and efficient, why aren’t they used more often in real-time weather forecasting? According to Samala, one reason for their rarity is the heavy computing requirements. “One significant factor is that coupled models require substantially more computing power, with dedicated high-performance computation needed to run a coupled model with the necessary very high spatial and temporal resolution.”
Making waves in forecasting
According to Samala, after extensive testing and validation, the WaveWatch 3 framework is now running operationally in “three nested domain set-ups” on the Met Éireann high-performance computing server.
As a result, a two-way coupled model of Harmonie-Arome and WaveWatch 3 has been successfully implemented to predict wave parameters and winds over Ireland and the surrounding ocean, while the ROMS standalone ocean model has been tested for ocean state forecasting.
“This coupled (atmosphere-wave) forecasting system is the first of its kind for Ireland and will have significant implications for renewable energy applications by providing improved weather forecasts and an integrated solar, wind (onshore and offshore) and wave energy forecasting system,” says Samala. “The coupled model will result in better forecasts for weather and ocean state, and in turn, for renewable energy forecasts on the desired time scales and heights.
“The next step is to couple this model with an ocean model, resulting in a fully coupled ocean-atmosphere-wave model.”
From wind to wave
So, how does this coupling work?
Samala says that due to the complexity and computational intensity of weather and ocean models, they are typically run separately with “prescribed boundary conditions”.
“The atmosphere model passes wind forecasts in real time to the wave model, which in turn passes the sea surface roughness back to the atmospheric model,” he explains. “Validation experiments showed that the coupled model forecast improved compared to the standalone models. Wind and wave forecasts were compared for 48-hour forecasts with available observations.”
The complexity of these models – particularly the ocean and atmosphere models – has proven to be a challenge, according to Samala.
“Understanding these complex models and coupling them using a suitable coupler in real time is a challenging task,” he says. “These models require high-performance computers to run, so dependency on computer architecture is another challenge.
“A dedicated group of people working together with continuous interaction with the developer community is the best way to handle these challenges.”
Against the backdrop of a worrying few years of weather extremes, from record-breaking temperatures to disastrous levels of precipitation, an improved weather forecasting system is an essential component of our preparedness for future climate difficulties.
Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.
