Geoscience Reference
In-Depth Information
6. Improved Use of Observations—Hurricane Forecast
and Improvement Project (HFIP)
In order to increase the pace of forecast improvement NOAA developed the
Hurricane Forecast Improvement Project (HFIP) whose goal is to decrease the
error in forecast guidance for track and intensity by 20% in five years and 50%
in 10 years (HFIP Plan). In order to achieve such an ambitious goal HFIP is
pursuing a strategy that is focussed on making use of the aircraft observations
to provide better initial analyses of the storm's structure for the models to start
from. Recent results from the programme show that assimilation of aircraft
observations, particularly airborne Doppler winds shows promise of reaching
HFIP intensity goals (Gall et al., 2013).
Two research groups, one at the Pennsylvania State University using the
ARW model and working with HRD (Zhang et al., 2011a), and the other at
HRD using the HWRF model (Aksoy et al., 2012) demonstrated that
assimilating data collected by the NOAA P-3 tail-Doppler radar in hurricanes
during the annual Intensity Forecast Experiment (IFEX) (Rogers et al., 2006)
run by HRD, as well as other data collected by the P-3 and Air Force C-130
hurricane hunter aircraft (including flight level and dropsonde data) can
significantly improve forecasts of intensity. The results compared to the 5-Day
Statistical Hurricane Intensity Forecast Model (SHIFOR5) are shown in Fig. 9
for most cases for which radar was taken in storms during 2008-2010.
Note that for all forecast lead times from 36 to 96 hours inclusion of aircraft
data improves the forecasts by 30% and as much as 50% over SHIFOR5 and
for those lead times are close to or exceed the 20% HFIP goal and approach the
50% goal for these cases. This is the first time that dynamical models have
shown this level of skill for intensity forecasts. It is likely that this result is
because the high-resolution data near the core better defines the environment
flow and storm structure within the hurricane and hence the impact, even out
to five days. These analyses avoid the use of filtering and bogusing to create
initial conditions that are used in most regional hurricane models. Note also
that there is a problem in the early part of the forecast where the model is still
adjusting to the initial conditions. Fixing this initial problem is a major current
focus of HFIP.
Note that even the HWRF model without adding aircraft data (red line in
the figure) gives a substantial improvement relative to the baseline. This
improvement suggests that the newer regional models (HWRF and ARW) could
provide an improvement as compared to the HFIP baseline (likely because of
higher resolution) and that adding aircraft data, when available, could add
another 10-20%. This would meet (at some forecast periods) the HFIP interim
(five-year) goal of 20% improvement, for those initializations for which tail
Doppler radar and other aircraft data are collected. For other times, HFIP is
pursuing the use of high-resolution satellite data near the hurricane.
Search WWH ::




Custom Search