 |
| Titel |
Improving Hydrological Models by Applying Air Mass Boundary Identification in a Precipitation Phase Determination Scheme |
| VerfasserIn |
James Feiccabrino, Angela Lundberg, Nils Sandström |
| Konferenz |
EGU General Assembly 2013
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 15 (2013) |
| Datensatznummer |
250083914
|
|
|
|
|
|
| Zusammenfassung |
| Many hydrological models determine precipitation phase using surface weather station data.
However, there are a declining number of augmented weather stations reporting manually
observed precipitation phases, and a large number of automated observing systems (AOS)
which do not report precipitation phase. Automated precipitation phase determination suffers
from low accuracy in the precipitation phase transition zone (PPTZ), i.e. temperature range
-1Ë C to 5Ë C where rain, snow and mixed precipitation is possible. Therefore, it is valuable
to revisit surface based precipitation phase determination schemes (PPDS) while manual
verification is still widely available.
Hydrological and meteorological approaches to PPDS are vastly different. Most
hydrological models apply surface meteorological data into one of two main PPDS
approaches. The first is a single rain/snow threshold temperature (TRS), the second uses a
formula to describe how mixed precipitation phase changes between the threshold
temperatures TS (below this temperature all precipitation is considered snow) and TR (above
this temperature all precipitation is considered rain). However, both approaches
ignore the effect of lower tropospheric conditions on surface precipitation phase. An
alternative could be to apply a meteorological approach in a hydrological model.
Many meteorological approaches rely on weather balloon data to determine initial
precipitation phase, and latent heat transfer for the melting or freezing of precipitation
falling through the lower troposphere. These approaches can improve hydrological
PPDS, but would require additional input data. Therefore, it would be beneficial
to link expected lower tropospheric conditions to AOS data already used by the
model.
In a single air mass, rising air can be assumed to cool at a steady rate due to a decrease in
atmospheric pressure. When two air masses meet, warm air is forced to ascend the more
dense cold air. This causes a thin sharp warming (frontal inversion) of air in the vertical
profile between the lower cold air mass and the warm air mass above. The warm air
forced up often cools to its condensation temperature, becoming the main cause of
winter precipitation. A common exception comes with cold air mass boundaries
(CAMB) not having a frontal inversion in the vertical profile. Therefore, CAMB
precipitation occurs under very different lower tropospheric conditions, than other
precipitation.
Changes in continuous hourly AOS temperature and wind could be used to identify
different types of surface air mass boundaries. When identified rain and snow observations
occurring immediately before CAMB were separated from all other observations, the TS and
TR values -1Ë C, 3Ë C respectively, were found to be 1Ë C cooler than the TS and TR for
non-CAMB observations. Analyzing CAMB separately reduced total misclassified
precipitation from 7.0% to 5.4% (23% improvement) in the PPTZ. However, this tool only
allows a statistically better chance for correct precipitation phase determination; it is
incapable of adjustments for deviations from an average vertical temperature lapse rate. |
|
|
|
|
|
|