Pitfalls of global geoscientific analysis of river deltas based on Big Data

This is a Nature Matters Arising submission to the Nienhuis et al., 2020 paper: Nienhuis, J. H., Ashton, A. D., Edmonds, D. A., Hoitink, A. J. F., Kettner, A. J., Rowland, J. C., & Törnqvist, T. E. (2020). Global-scale human impact on delta morphology has led to net land area gain. Nature, 577(7791), 514-518.

Abstract:

With the advent of Big Data, geoscientists are confronting global statistics to derive universal drivers of various earth systems^1–3. Nienhuis et al.¹ identified 10,848 deltas and retrieved fluvial, wave and tidal sediment fluxes from global models and estimations, based on simplified equations and Aqua Monitor(Aqua)⁴ and Global Surface Water Explorer(GSW)⁵ datasets, to obtain surface area change for each delta. They found that, globally, deltas have gained 54 ±11.8 km²/yr over the last 30 years due partially to human interventions in drainage basins. This result is not surprising, inasmuch as deltas grow naturally by accumulating sediments, and confirms a previous study⁶. Nienhuis et al.¹ attributed land loss in about 1000 deltas to recent reductions in sediment supply.

Identifying globally 11,000 deltas is significant at first sight, given the current concerns with delta vulnerability^7–11, backdrop of the study by Nienhuis et al. Prompted by the inventory of numerous deltas in regions such as the British Isles and Britanny, France, we ran a check on the Nienhuis et al. results by randomly selecting 108 deltas from their dataset (n = 1%; Supplementary Methods) and obtained no agreement with the original data¹. We raise concerns about their methodology and consider that their main result, namely global delta area change attributed to human impact is unsubstantiated because drawn from a database with errors and statistics that could not be validated. These include false-positive delta detection, overlapping buffers, non-robust statistics, and data display.