 |
| Titel |
Assessing uncertainties in superficial water provision by different bootstrap-based techniques |
| VerfasserIn |
Dulce B. B. Rodrigues, Hoshin V. Gupta, Eduardo Mario Mendiondo |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250099993
|
| Publikation (Nr.) |
 EGU/EGU2014-15856.pdf |
|
|
|
|
|
| Zusammenfassung |
| An assessment of water security can incorporate several water-related concepts,
characterizing the interactions between societal needs, ecosystem functioning, and
hydro-climatic conditions. The superficial freshwater provision level depends on the methods
chosen for “Environmental Flow Requirement” estimations, which integrate the
sources of uncertainty in the understanding of how water-related threats to aquatic
ecosystem security arise. Here, we develop an uncertainty assessment of superficial
freshwater provision based on different bootstrap techniques (non-parametric resampling
with replacement). To illustrate this approach, we use an agricultural basin (291
km2) within the Cantareira water supply system in Brazil monitored by one daily
streamflow gage (24-year period). The original streamflow time series has been
randomly resampled for different times or sample sizes (N = 500; -¦; 1000), then
applied to the conventional bootstrap approach and variations of this method, such as:
“nearest neighbor bootstrap”; and “moving blocks bootstrap”. We have analyzed
the impact of the sampling uncertainty on five Environmental Flow Requirement
methods, based on: flow duration curves or probability of exceedance (Q90%, Q75% and
Q50%); 7-day 10-year low-flow statistic (Q7,10); and presumptive standard (80% of
the natural monthly mean ?ow). The bootstrap technique has been also used to
compare those “Environmental Flow Requirement” (EFR) methods among themselves,
considering the difference between the bootstrap estimates and the "true" EFR
characteristic, which has been computed averaging the EFR values of the five methods and
using the entire streamflow record at monitoring station. This study evaluates the
bootstrapping strategies, the representativeness of streamflow series for EFR estimates and
their confidence intervals, in addition to overview of the performance differences
between the EFR methods. The uncertainties arisen during EFR methods assessment
will be propagated through water security indicators referring to water scarcity
and vulnerability, seeking to provide meaningful support to end-users and water
managers facing the incorporation of uncertainties in the decision making process. |
|
|
|
|
|
|