City #9: Duluth, MN (KDLH)

The WxChallenge moves on from Newark, NJ to Duluth, MN. We are forecasting for Duluth (KDLH) from 11 March - 21 March for the 24-hour period 06 UTC - 06 UTC which translates to 1 AM CDT if you are looking at NWS Duluth, MN. Forecasts are still due by 00 UTC which is 8 PM local time (note the time change due to Daylight Savings Time beginning).

Duluth is located in Eastern Minnesota on the shore of Lake Superior. Now one might think that since it is located on a Great Lake that it is prone to lake effect. While this is true, there is another subtlety about Duluth. According to its Wikipedia page, it is the "San Francisco of the Mid-West" due to the proximity to a body of water and sloping terrain. So we have a combination of lake-effect and topographical influences to worry about given the right flow setup as I'll discuss later! That's pretty interesting, in my opinion.  Taking this into account and stalking the airport (KDLH WMO ID = 72745) we see that the portion of the city that borders the Lake is around 185 meters high, however KDLH is located to the northwest of the shore of Superior ( ~10 km away) and is located at an elevation of 435 meters.


View Larger Map

Given the orientation of the lake (southwest -> northeast) near the city and the fact that the land abruptly slopes up in elevation at the shore, upslope lake-effect or lake-enhanced precipitation may occur with easterly or northeasterly flow from synoptic-scale systems passing to the south and east. I wanted to find out the temperature of the lake, via the Great Lakes Observing System but the data are not available either due to the buoys being taken out for the winter or the lake froze over. Anyone have any clue which hypothesis is correct?

Large-scale Pattern:
A look at forecast 500-hPa heights according to the GFS shows that a trough will remain in place for the majority of the first forecast day, with a shortwave embedded within the flow towards the end of this week. This trough should persist until next week with generally dry and chilly air being advected by the wind from Canada. Next week a ridge builds out ahead of a developing cyclone lee of the Rockies which may allow for some above-average temperatures. This is all just from glancing at one run of one deterministic model so take it with a grain of salt... or the whole salt shaker depending on your outlook on life.


Tips/Comments for Duluth, MN (KDLH):

• Look at more than one deterministic model, especially side-by-side comparisons, toggling between models (Mode: Clickable Map), or even better- ensembles (SREF/GEFS Plumes, SREF Maps, GEFS Maps).
• WxChallenge's wind forecast is for sustained, not gusts. Instead of just looking at plan-view maps at or near the surface, try taking a look at forecast soundings. The measured wind speed is a 2-minute average. Gusts often occur because with heavy precip, higher momentum air can be brought down to the surface, i.e. winds reach the surface via more than just mixing, but unless they persist for about 2-minutes they won't verify. 
• Upslope lake-enhancement of precip may occur given easterly or northeasterly flow, the High Resolution Rapid Refresh (HRRR) is run out only for 15 hours from the initialization time, but they do have a "Great Lakes" domain option if you wanted to take a look at a high-resolution product if you had some questions about the very near-term part of your forecast.
• Really read the NWS Area Forecast Discussion (AFD). Those forecasters are experts in the intricacies of forecasting for Duluth.

 Tips/Comments for Forecasting in General:

• Check out the current weather first via satellite, radar, surface observations, METAR, etc. You have to understand the present before you can look to the future.
• Get a feel for general pattern from NAM, GFS, etc. (available from here and here, for example). Identify the general pattern of low and high surface pressure systems and their progression, temperature advection at 850 mb, moisture at 700 mb, vorticity at 500 mb and the general jet structure at 300 mb. Make sure you are focusing your attention on Duluth.
• Look at forecast model soundings for the period for KDLH.
• Compare your idea for values with MOS (model output statistics).  Keep in mind these are instantaneous output values and you can always have values above/below that between output times. 
• Read the Duluth, MN WFO forecast discussion for KDLH.

Tips/Comments for Forecasting for High/Low temperature (Fahrenheit):

• Recall such basics as nighttime radiative cooling and daytime solar radiative heating, both of which are affected by cloud cover and fog.
• Frontal passages can affect your temperature forecast. Pay attention to the forecast timing of such frontal passages. 

Tips/Comments for Forecasting Wind Speed (knots):

• NWS outputs wind speed in mph so if you are comparing your forecast value to theirs keep that in mind. Otherwise the models output wind speed (WSP) in knots.
• Look through METAR or surface obs for wind direction and temperature changes.

 
Tips/Comments for Forecasting Precipitation (inches):

• Take a look at some higher-resolution models (e.g. WRF)
• What does the WPC (formerly known as HPC) show?
• Take a look at some ensemble data (e.g. SREF (if no local option then try SPC's SREF page))
• Most products you will be looking at will output the QPF in inches already so no conversion is necessary.

I don't know much about Duluth except for the fact that Bob Dylan was born there. So how do we successfully forecast for Duluth? The answer, my friend, is (quite literally) blowin' in the wind. If you didn't get that, culture yourself. Go Team SBU!!