 |
| Titel |
Spatial organisation in hydrological model structure for New Zealand catchments |
| VerfasserIn |
Hilary McMillan, Ross Woods, Martyn Clark |
| Konferenz |
EGU General Assembly 2013
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 15 (2013) |
| Datensatznummer |
250077636
|
|
|
|
|
|
| Zusammenfassung |
| Hydrologists increasingly agree that a single hydrological model structure is unlikely to be
suitable for all catchments: instead, models should be selected according to characteristics of
the catchment. Our challenge is to determine how to select the most appropriate
model structure. This complex question requires that we use observed data to infer
dominant runoff generation processes, and translate this process knowledge into
model structure choices. We can then ask questions such as: over what scales do
recommended model structures change? How much data is needed to select model
structure? How can we generalise model structure choices to catchments where data is
scarce?
In this presentation we address these questions, using the New Zealand landscape as our
‘virtual laboratory’. New Zealand is an excellent location to test hypotheses relating to
model structure, due to its rich diversity of hydrological landscapes. Landscape
types range from temperate rainforest with steep, bedrock gorges, through rolling
pasture, to alluvial plains with braided rivers. Our method is to apply diagnostic
signatures, which use a range of hydrological data types, to target specific aspects
of model structure choice. We bring together results from national hydrometric
networks, and in-depth studies in experimental catchments, to explore organisation,
similarity and diversity in recommended model structures across the New Zealand
landscape.
To identify model structures which are consistent with measured data, we use a range of
diagnostic signatures tailored to the data types available. At the national scale, networks of
rain and flow gauges are used to investigate runoff ratio, recession characteristics and
threshold responses to precipitation and soil moisture. At the experimental Mahurangi
catchments, dense networks of 13 rain, 27 flow and 36 soil moisture gauges within a 50 km2
area enable us to evaluate small-scale patterns and diversities of model structure. In contrast,
the experimental Waipara catchment in the Eastern foothills of the NZ alps provides
networks of 20 soil moisture sensors and 10 shallow groundwater wells within a 1 km2
catchment, as well as deep groundwater wells and 5 nested flow gauges. This data
enables us to test for additional aspects of model design related to groundwater
response.
We relate the local responses and diagnostic signatures to the wider, national-scale
patterns. We consider whether local and national model recommendations are compatible,
and how model structure patterns and diversity change with scale. Finally, we consider how
uncertainty in measured data sources in NZ has the potential to affect diagnostics and
hypothesis testing for model structure. |
|
|
|
|
|
|