Growing numbers of citizen scientists are joining the professional community to map structural damage caused by natural disasters such as Typhoon Haiyan, which struck the Philippines in 2013 (see Nature 515, 321; 2014). However, stakeholders need to recognize some crucial limitations of crowdsourced contributions.

Citizen scientists are invaluable for mapping roads accurately and for picking out local points of interest. But when it comes to analysing structural damage, even high-resolution satellite data are inadequate. For example, fine-resolution aerial imagery revealed that damage after the 2010 Haiti earthquake was ten times worse than originally concluded by mapping volunteers. Moreover, vertical data cannot provide a comprehensive picture, particularly of partial damage (see N. Kerle and R. R. Hoffman Nat. Hazards Earth Syst. Sci. 13, 97–113; 2013).

Furthermore, volunteers receive only minimal instruction. They are not trained to map structural damage, which relies on image proxies such as shadows and debris, nor do organizers of the mapping systems provide corrective feedback for improving mapping accuracy.

Coordination with professional organizations on standards and terminology must also be improved if the needs of disaster-response stakeholders are to be met.