Share
According to the European Space Agency (ESA), the new constellation of Sentinel satellites will deliver 8TBs of freely available raw image data every day by 2020. This is just the tip of the iceberg when other space programmes and data products derived from the raw data are considered. As a colleague at a recent conference pointed out, âItâs like standing next to a dripping tap, expecting it to suddenly turn into a gushing hosepipe.ââ©
The potential applications for this free and relatively accessible data are numerous, ranging across the fields of conservation, development, economic strategy and agriculture. The beneficiaries and potential users range from government institutions, small, medium and large businesses to a raft of different civil society organisations and NGOs.â©
So, how are we going to cope with these data volumes and ensure that we can extract useful information? Technology companies, including ours, get excited about new ways of handling large, continuous flows of data, extracting arrays of functions for complex polygons for particular timeframes, and then processing these into user-accessible graphics and tables. â©
Technology is undoubtedly an important part of the solution, but the main issues holding back the exploitation of this Big Data treasure trove are social and economic. In the absence of a clear understanding of what end user content is required and without clear business cases for delivery, the technology on its own will not deliver viable services. â©
The EO community is therefore grappling with the problem of how to turn vast resources of freely available data into economically viable information products.â©
A regional solutionâ©
Initial applications of these EO labs in Brazil and Mexico are likely to be in assessing the risks to different forest ecosystems from agricultural expansion, human induced fires and extraction of timber for fuel and construction. While many projects in both countries have sought to avoid deforestation and degradation the measurement of outcomes these efforts in terms of forest retention and regrowth have yet to be carried out systematically, in a way that informs the design and implementation of project interventions, and guides policies promoting sustainable agriculture and bioenergy.â©
A challenging taskâ©
Long-term viabilityâ©
There is a common perception among end users that EO products should be free and open source. While that may be desirable for civil society, there is a need to understand how such products will be paid for. Data processing, classification, calibration and quality controls all require time, effort and resources that are not readily provided by an open data model in the absence of a benevolent state or multinational IT company. We therefore need to consider models of a keystone client for a given information product and then tiers of free and paid content for users requiring different levels of detail.â©
Our aim in the short-term is to define a core set of good quality, regularly updated, EO products that can be sustained in the long-term in each country. The core can then be expanded on and improved over time as the value added by the service becomes apparent.â©
We think that the EO lab model should prove to be a flexible and cost effective delivery mechanism, since it uses scalable cloud computing architecture and can be set up in an existing research organisation without the need for new physical infrastructure. Of course, technology is still important and new advances in the handling of Big Spatial Data allow complex queries to be run on compressed, stacked, flat files without the need to unpack or generate tables.â©
However, the greatest effort will likely be required in the less glamorous area of dialogue with multiple stakeholders, business planning and collaboration. The sustainability of Big EO Data will come down to good, old-fashioned business sense.â©
The sustainability of Big EO Data will come down to good, old-fashioned business senseâ©
Richard Tipper is chairman of Ecometrica (www.ecometrica.com)
