4.2 NUMERICAL WEATHER PREDICTION FOR FIRE HAZARDS ON OAHU/HAWAII
Duane E. Stevens*
and Derek Funayama
University of Hawaii,
Honolulu, Hawaii
Francis Fujioka
USDA Forest Service,
Riverside, California
The challenge of
weather forecasting in Hawaii has been dominated by two frontier problems:
hurricanes and severe weather. Severe weather issues include high wind
events (usually marine) and flash floods, both mesoscale features.
Brush fires, especially during the summer months, are hazards that occasionally threaten the leeward (Waianae) coast of Oahu in Hawaii. The prediction of future surface wind patterns is crucial for determining whether brush fires are to become a hazard to life and property. We use the mesoscale spectral model (MSM) developed at the National Centers for Environmental Prediction (NCEP) of the National Weather Service (NWS), with horizontal resolution of 2 km, to address the problem of forecasting the vulnerability of fire-prone regions under various weather conditions.
In the complex terrain of Oahu, the MSM provides a prognosis of surface parameters which can then be used as input to predict the spread of brush fires. High temperature and low humidity are indicative of areas which are prone to fire. Wind direction and speed information suggest areas where fire is likely to propagate once it starts. Accurate prediction of precipitation, if viable, would greatly aid mitigation efforts. The relative strengths of local trade winds and the diurnally varying sea breeze -- land breeze circulations are likely to significantly influence evolving fire hazards. In 1998 concern was raised over the military use of the environmentally-sensitive Makua Valley on the leeward coast of Oahu. Over a ten year period, approximately 300 fires had been initiated during live-fire training exercises. These fires threaten a detrimental impact on environmentally protected and potentially environmentally protected species (Honolulu Star-Bulletin, September 28, 1999). A local environmental group threatened legal action against the Army to prevent future military use. Subsequently, the Army temporarily ceased its training activities, which policy continues to date.
As part of its response
to this environmental challenge, the Army is developing a plan to mitigate
the deleterious effects of live-fire training in Makua Valley. Circulation
features on the scale of this valley are being investigated under the auspices
of a cooperative agreement between the USDA Forest Service and the University
of Hawaii at Manoa.
Figure 1 displays
the smoothed orography of the western region of the island of Oahu, most
populous in the State of Hawaii. The western leeward coast is dominated
by the Waianae Valley. Between Wainae and western-most Kaena Point lies
Makua Valley, bounded by the ocean (southwest) and Waianae Range (northeast),
along with two lateral ridges. The dimensions of Makua Valley are approximately
3 km by 5 km. Unfortunately, Makua Valley is essentially missing from the
smoothed orography presently being used.
The Hawaii Weather/Climate
Modeling Ohana (HWCMO) makes a 24-hour forecast each day for the Hawaiian
archipelago, Maui, and Oahu using the NCEP Mesoscale Spectral Model (Juang,
1992). Time-dependent lateral boundary conditions are prescribed from the
NCEP 00Z Aviation Model forecast. These three non-hydrostatic model runs
are executed with horizontal resolution of 10 km, 3 km, and 2 km, respectively.
Results are published daily at http://www.mhpcc.edu/~wswx.
Brush fires, especially
during the summer months, are hazards that impact the leeward (Waianae)
coast of Oahu, Hawaii with much of this region covered by grass and brush.
The island itself consists of two mountain ranges (Koolau and Waianae ranges)
that are almost perpendicular to the prevalent trades. The showers associated
with the flow encounters the first of the two ranges leaving little precipitation
to carry over to the leeward coast. These hot and dry conditions turn the
vegetation into good fuel sources for brush fires. Since this is an island,
residential areas are often located nearby, so that fire propagation is
an important issue.
The use of the mesoscale spectral model (MSM) at a high horizontal resolution (2km) addresses the problem of forecasting fire-prone regions under various conditions. The MSM provides useful parameters to predict the probability of occurrence of and the spread of brush fires. Temperature and moisture fields indicate tendencies of potential fire-hazard regions. Surface wind patterns become crucial for predicting if brush fires will become a hazard to life and property. Predicting precipitation assists with analysis of possible mitigation of fire hazards.
Two case studies are used to address trade wind conditions. Such conditions can be classified into two types: weak and strong. The weak case has winds under 5m/s with a sea breeze circulation often existing on the leeward coast. The strong case has winds that can exceed 10m/s. While May is not strictly a summer month, it serves to provide two days when weak and strong trades were present. May 22 and May 25 1999, the weak and strong trade winds respectively, were chosen to study the leeward coast using time series and spatial analyses. The variables used for the study are surface temperature and relative humidity, along with 10-meter winds and precipitation. Time series analyses are performed starting from 5AM to 2PM local time. Hence daytime conditions are addressed from sunrise to the afternoon.
The May 22 (weak trade wind) case shown in Figure 1 indicates a locally reversing sea breeze circulation. Model precipitation develops with the sea breeze.
On May 25, strong trades overcome any surface heating with flow relatively unhindered by the island of Oahu, as shown in Figure 2. Precipitation also occurs mainly due to topographic effects. Streamline analysis illustrates flow patterns for both days.
The analyses for both
cases are based on MSM data using model topography. This is helpful in
analyzing meteorological phenomena at the two-kilometer scale. The topography
shows a smoothed terrain interpolated from actual data. Despite a resolution
of 2 km, modeling based on such smoothed topography is of qualitative value
only. Specifically, entire valleys several kilometers wide are not represented.
Higher resolution will be required to pinpoint the source and spread of
brushfires over smaller scales.
While the smoothed
topography is capable of distinguishing between sea-breeze-type circulations
on weak trade wind days and the predominant surface easterlies on strong
trade wind days, more grid points are required to resolve the smaller (and
more typical for Oahu) valleys such as Makua. We are making plans to reduce
the grid size to sub-kilometer scale, which would greatly increase the
computational burden. In addition, interpolation schemes are being considered
as a simpler alternative to higher resolution "full physics" numerical
models.
This work is supported
by a cooperative agreement between the US Department of Agriculture's Forest
Service and the University of Hawaii at Manoa. Computing is performed at
the Maui High Performance Computing Center.
Juang, H.-M., 1992: A spectral fully compressible nonhydrostatic mesoscale model in hydrostatic sigma coordinates: formulation and preliminary results. Met. Atmos. Phys. 50, 75-88____________________________________________
Fig. 1. The May 22, 1999 case features a two kilometer modeled sea breeze circulation on the leeward coast of Oahu, Hawaii. Heights, displayed in meters, are modeled showing significant smoothing and height reduction of the topography. At this level, the main (Waianae) valley on the leeward coast can be seen. Note that the wind field is scaled such that one full barb indicates one meter per second and precipitation is in millimeters.
Fig. 2. The
May 25, 1999 case features strong trade wind flow masking any thermal effects
caused by surface heating or other mechanisms. Precipitation seen are topography
induced or by offshore convergence. All units are the same as in Fig. 1.