Downdraft Convective Available Potential Energy (DCAPE)
Meaning, calculation and interpretation of DCAPE
Quasi-Daily PNW Maps - Experimental, basis not yet peer-reviewed
2 July 2010: I will post near-daily updates to the observed DCAPE time series here, as well as occasional time-series maps.
New for 2010: Daily maps of DCAPE across the PNW are available here.
DCAPE, CAPE, and KE time series for UW WRF 12 km grid initialized 26 August 2010 00 UTC.
The patterns for Baker and Burns suggest convective and wind energy will both be high, and DCAPE will drop today. Hehe Butte is not overly worrying, though the winds will be somewhat high.
Nampa and Twin Falls, ID, flanking the Long Butte fire, both have seemingly benign wind energy (KE) today, somewhat higher tomorrow. The basic CAPE is also quite low for both locations. However, the dashed lines indicate KE and CAPE for surface air modified by 2C and they show that with a bit of heating, convection could rise quite high and mix downward significant winds both today and tomorrow.
Explanation
Evaporation to bring initally unsaturated air to saturation cools the air to its wet bulb temperature. If that air descends and evaporation continues to hold it at saturation, it will maintain a constant wet bulb potential temperature, warming slightly. On a skew-T chart, the air parcel would descend on a saturated adiabat. As long as the parcel is cooler than its environment during descent to the ground, it is denser and therefore has negative buoyancy. The Downdraft Convective Available Potential Energy, DCAPE, is a measure of the total negative buoyancy and the energy that a sinking, saturated air mass would have when it reached the ground.
The figure below shows a graphical example of DCAPE using a skew-T chart. The red and blue lines show the potential temperature and moisture, respectively, of air that starts at 750 mb and rises until saturated. The intersection of the two lines near 400 mb indicates saturation, and the green line from this point to the bottom of the chart indicates the line of constant wet bulb potential temperature for the air to subsequently descend to the ground. The brown shaded area is proportional to DCAPE. The graphical process is analogous to the way CAPE is often depicted on a skew-T diagram.
DCAPE was originally formulated by Emanuel (1994) for application to downdrafts in thunderstorms. As an outgrowth of a broader study of atmospheric energy and fire danger, I have seen indications that DCAPE shows a distinct pattern at the time of a significant runs on a number of historic fires. The same pattern seems to match subjective identification of upper level ridge breakdown in the Northwest, and a very similar pattern seems to accompany thermal troughs. These associations are all tentative and I've just begun to test them on other fires, so it is extremely important to consider this an experimental tool.
Calculating DCAPE requires selecting a starting height for the air under consideration. I have been using 3000 m AGL. I looked at DCAPE for 2000 m and 4000 m AGL for several cases, and found that the strength of the signal was weaker at these heights - I do not know why. Figure 2 shows a trace of DCAPE for a subjectively identified ridge breakdown in eastern Oregon in August 2008. Data from the University of Washington WRF 12 km grid was used for these calculations. The ridge built from August 13, through the 17th. Beginning on the 17th and continuing through the 19th, the ridge broke and lower pressure moved into the area.
As I mentioned above, I have seen similar patterns in DCAPE for large fires. Below is the DCAPE time series for the Ham Lake fire in northern Minnesota in 2007. The fire made its biggest run on day 131, just as the DCAPE value was falling from about 600 J/kg to -250 J/kg.
A similar plot for the 2008 Rattle Fire in Oregon, showing both DCAPE and the area growth rate (in hectares per hour) shows a similar spike in growth at the time of the DCAPE drop. In this case, the drop in DCAPE went from nearly 1400 J/kg to about 300 J/kg and continued for 5 days.
