SPM fluxes estimates from discrete monitoring : comparison of calculation methods and uncertainties
Résumé
The SPM (suspended particulate matter) fluxes temporal variabilities are even more complex than water flows and are therefore very difficult to estimate. In France, SPM are surveyed within the national water quality survey system (RNB) and are monitored with a bimonthly or monthly frequency, as other water quality indicators. The flux calculation methods and their related uncertainties on the basis of an experimental data set of continuous – i.e. daily monitoring for contrasted river basins ranging from 100 to 100 000 km2 (in France and in USA, 60 stations totalling 600 station-year of daily SPM and flows). The uncertainties, biases, and errors, result from the analysis of the distribution of errors made on simulated surveys at various frequencies (3, 5, ... 30 days) in comparison to the same value. The duration curves of SPM fluxes and the key indicator M2 %, i.e. the percentage of SPM carried in 2 % of time on a multiannual period (4 to 20 y), are then determined. The M2 % indicator is used to build-up a nomograph linking biases (median or errors) and imprecision (difference between upper and lower deciles of errors) to monitoring frequencies (3, 4, ... 30 days). This nomograph can be used to : i) optimize surveys frequencies, given a flux error target, or to ii) associate error bars. The nomograph is so far established for the mean discharge-weighted concentration method, one of the most common method with the rating curve method.