Armenian weather predictions get help from the Cloud

Extreme weather made headlines again last year. 2018 saw frozen iguanas fall from the trees in Florida, wildfires devastate parts of Sweden, Greece and the US, and snow fall in the Sahara. So, when weather conditions can turn perilous so quickly, any advances in short-term forecasting can potentially make a huge impact.

Clouds of the IT kind are working to give Armenian scientists the upper hand against the extreme weather events that have cost the country around 177 million USD between 2009 and 2013. It’s all thanks to the Infrastructure-as-a-Service (IaaS) cloud computing solution, that has been deployed by ASNET-AM, the Academic Scientific Research Computer Network of Armenia, following a 2017 workshop organised by the EaPConnect project.

ASNET-AM – established 25 years ago by IIAP NAS RA (the Institute for Informatics and Automation Problems of the National Academy of Sciences of the Republic of Armenia) – is currently being used by more than 65 organisations and the country’s scientific community.

Dr. Artur Gevorgyan (pictured), head of Division of Development and Validation of Hydrometeorological Models, Service of Hydrometeorology and Active Influence on Atmospheric Phenomena at the Hydromet Service of Armenia believes that IaaS can remove some of the obstacles that have prevented forecasting in the past.

“Strong winds, storms, heavy rain, heatwaves, and severe frosts all cause hundreds of millions of dollars of damage in the region. Also we had loss of lives and injuries so it is a very important issue,” he explained. “More accurate ‘nowcasting’ and predictions of severe weather events by numerical weather prediction models could reduce that.”

‘Nowcasting’ uses surface weather station data, wind profiler data, and any other weather data available to forecast for a period of 0 to 6 hours. In this time range it is possible to forecast small features such as individual storms with reasonable accuracy.

This hasn’t been possible in the past, because high-resolution computer modeling capable of forecasting over Armenia’s mountainous terrain has not been available. Dr Artur Gevorgyan explained: “These models have not been used operationally at the Hydromet Service of Armenia because of the absence of High-Performance Computing (HPC) resources. The IaaS cloud service can address this, as it can be used for HPC tasks.”

ASNET-AM’s IaaS solution

So how has ASNET-AM harnessed the cloud? Following a workshop in June 2017, organised by the Eastern Partnership Connect (EaPConnect) project, ASNET-AM staff developed a working relationship with GARR, the ultra-broadband network for the Italian research and education community. The team behind the Armenian project worked with GARR to deploy the experimental cloud IaaS infrastructure to use OpenStack middleware. The result is the fast-growing, cost-effective, open, flexible and scalable cloud management platform.

The benefits have been immediate. Those using it no longer have to ensure uptime or maintain hardware and networking equipment. Some services have already been migrated to the cloud platform, such as network monitoring, and IaaS-customised Virtual Machines (VMs) are now being provided to the end users. This means for example that librarians can use several VMs to work with different digital library environments.

Completely free for ASNET-AM end users, the cloud service is expected to increase the sustainability of the NREN, as well as providing support to societally important research, such as weather forecasting.

“The results are very promising,” said Dr Gevorgyan. “Now we are able to run sophisticated models for research purposes. We’ve studied heavy rains, strong wind, and we are now studying hail storms.”

“Before this collaboration, nothing had been done on modelling weather using numerical weather prediction models for Armenia.”

“It is a very new job, but I think in future we will continue to see progress. It’s quite simply a new era.”