Out of Africa

“If we are going to do anything more effective and meaningful than small scale pilots of climate services then as a sector we have to work on data standardisation and hence, interoperability”

Posted January 2019

By Prof David WallomAssociate Professor and Associate Director – Innovation of the Oxford e-Research Centre

It’s not unusual in my line of work to come back from speaking at a data research event laden down with fresh insights, new ideas and a fistful of business cards, but even by my standards the ‘to do’ list I brought back from International Data Week in Botswana is daunting.

I am on a mission to promote the improved standardisation of data around the world.

I presented a paper ‘Using big data to better understand climate change impact using Climateprediction.net’ (David C.H. Wallom, Sarah N Sparrow & Neven Fuckar) in a session on November 6, at a side event, SciDataCon, which was chaired by Dr Mary-Jane Bopape, a Chief Scientist at the South African Weather Service.

With topics ranging from ‘Analysis of frequency and severity of droughts in Botswana’ and ‘Forest structure and ecological niche modelling’ to ‘Reducing the uncertainty of the African greenhouse gas budget’ in this session alone, it became increasingly obvious that there is a major obstacle. Many standards, particularly in climate data, are very broad and can be interpreted in many ways – all of which are actually compliant with existing standards but may not be interoperable with other implementations and this stands in the way of the progress of climate services out of academic research and into use by the public and private sectors.

One of the key things that emerged was that, in many cases where standard data is in use, there is a further need to profile the standards to cut down interpretation and, in turn, boost the potential for interoperability.

Lots of people acknowledge that this is an issue but there is very little effort going into solving it.

Data standardisation and interoperability

If we are going to do anything more effective and meaningful than small scale pilots of climate services then as a sector we have to work on data standardisation and hence interoperability. Only then will we enable the much-needed broadening out of climate service providers for the public sector, such as local authorities, and for the private sector so that citizens and organisations can make practical use of the climate services being developed.

An example of where benefit could be tangibly seen, atmospheric pollution came up in the context of African cities, consuming both Earth Observation and ground-based observation data, which has great potential for climate services which would be of use to both city authorities and to citizens. But to be meaningful it needs to be of a common standard.

Environmental data for climate services should always follow some kind of tightly defined profile rather than simply complying with a particular standard.

A good starting point would be identifying some example profiles of open data that are already out there and are recognised as usable by communities in Africa.

A real highlight of our session was that, other than Matt Fry of the Centre for Ecology and Hydrology and myself, every single other presenter was an African talking about a project they had developed and then run themselves.

There is a move across Europe to globalise our climate services research but this is not a question of imposing European research preconceptions.  It is too common when researchers from developed counties reach out to researchers in developing countries, to assume we have a good understanding of their specific, complex needs before finding out what those exact needs really are.

A new approach to data management training

I believe we need to take a holistic approach, starting with training and this came up frequently during International Data Week: training for data management, understanding best practice in data management and best practice around software services development.

Africa – thought to be the most vulnerable continent to climate change because of the reliance of its economy on weather and climate sensitive sectors – has very specific needs, it also has its own challenges around development-style projects.

So, what is top of my ‘to do’ list?

I am changing my approach to training researchers from developing countries. Whereas workshops I held for climate services researchers in Brazil and China were hosted here at the University of Oxford’s e-Research Centre, I am now planning with Dr Bopape a workshop in Africa in 2019 – perhaps returning to Botswana, a country of inspiring beauty and contrasts and a country where the passion and ambitions of its researchers have inspired me.

David Wallom, IEA big data mooc

Prof David Wallom is Associate Professor and Associate Director – Innovation of the Oxford e-Research Centre, where he leads two separate research groups Energy and Environmental ICT and Advanced e-infrastructure & Cloud Computing.