![Hier klicken, um den Treffer aus der Auswahl zu entfernen](images/unchecked.gif) |
Titel |
Optimisation of flux calculation in rivers from discrete water quality surveys, a step towards an expert system |
VerfasserIn |
S. Raymond, F. Moatar, M. Meybeck, V. Bustillo |
Konferenz |
EGU General Assembly 2009
|
Medientyp |
Artikel
|
Sprache |
Englisch
|
Digitales Dokument |
PDF |
Erschienen |
In: GRA - Volume 11 (2009) |
Datensatznummer |
250025450
|
|
|
|
Zusammenfassung |
Good estimates of fluxes of suspended particulate matter (SPM), total dissolved solids (TDS)
and nutrients and contaminants are required for both Earth System science and river basin
management. However, in most cases discrete sampling (weekly to monthly) is the rule. Few
flux calculation methods are commonly used, yet their performances, i.e. uncertainties for
given frequencies, at given stations and for each water quality variables, remain
unknown. Based on a rare set of 1085 station-year of daily flux record for SPM, TDS
and nutrients (dissolved and total), the performance of 9 calculations methods is
explored.
Discrete surveys at various frequencies (3days to 30 days) are simulated by Monte-Carlo
sorting (100 runs) on which the 9 fluxes are calculated (annual and interannual). At this stage,
the sub-daily variations of fluxes for the medium and large basins are not considered. The
dataset for SPM corresponds to 55 stations (600 to 600 000 km2 basin area), 34 stations (700
to 1000000 km2) for TDS and for nutrients we consider 9 stations for NO3-, NH4+,
PO43- and Ptot (600 to 30Â 000 km2). About 80% of the dataset originates from US
records (USGS and Lake Erie tributaries survey) and 20% from French stations, this
covering a wide range of hydrological and geochemical conditions in the temperate
zone.
Each sorted flux is compared to known fluxes established on daily records: percentiles of
their relative errors (e10, e50 and e90) are used to determine the biases (e50) and the
imprecisions (e90-e10) (Walling and Webb, 1981) which are then compared for each of the 6
water quality variables, for each flux methods and for various simulated survey
frequencies.
The calculation methods include 5 rating-curve approaches (linear“M1”, “M2”, Phillipps
et al, 1999) with and without Ferguson correction (Ferguson, 1987), polynomial, truncated at
discharges exceeding median annual or long-term water discharge), 2 methods based on
hydrograph separation (Phillips et al, 1999) including a quadratic runoff module (Bustillo,
2005), 1 linear interpolation method and 2 discharge-weighted concentration methods
(“M18”, “M19”, Philipps et al, 1999).
As expected, based on 55 stations and 430 years, SPM fluxes are the most uncertain ones
with maximum biases determined on annual fluxes (monthly sampling simulations).
ranging at stations from -75% to +55% by the classical rating-curves approach (“M1”,
“M2”) droping to -60% to +5% for the M18 method. At this frequency, biases are
much less for Ptot and PO4-3 (-30% to +10%), nitrate (-5% to +10%) and are
negligible for TDS. For higher frequencies, the biases are reduced: for instance for
weekly surveys they drop to -25% for SPM and to -20% to 5% for Ptot for the M18
method.
The river basin size is influencing the performance of calculations methods: SPM
flux errors are much higher for smaller basins (103 to 104 km2) than for larger
ones (> 104 km2), probably in relation with the flow duration in 2% of time which
is a key control factor of flux duration in 2% of time (Moatar et al, 2006). This
indicator based on daily flow (Q) records is generally available at water quality
stations. Other indicators based on discrete water quality surveys are being tested
to explain the performance of flux methods for each variable: concentrations (C)
variability, C vs Q relationship, concentration seasonality. For each variable and each
station the optimal flux calculation method will be derived from the future expert
system.
BUSTILLO V., Biogéochimie et hydroclimatologie appliquées à l’aménagement des
bassins fluviaux .PhD Thesis, INP Toulouse,232 p+annexes (2005).
FERGUSON R.I., Accuracy and precision of methods for estimating river loads. Earth
Surface Processes and Landforms, vol. 12,95-104 (1987).
MOATAR F., PERSON G., MEYBECK M., COYNEL A., ETCHEBER H., CROUZET
P., The influence of contrasting suspended particulate matter transport regimes on the bias
and precision of flux estimates Science of the Total Environment 370, 515-531
(2006).
PHILLIPS. J.M., WEBB B.W., WALLING D.E., LEEKS G.J.L., Estimating the
suspended sediment load of rivers in the LOIS study area using infrequent samples. Hydrol
Process, 13:1035-50 (1999).
WALLING D.E. et WEBB B.W., The reliability of suspended sediment load data IAHS
Publ., 133, 177-94 (1981). |
|
|
|
|
|