From a previous post you might remember that I found that by defining a surge through criteria of IWVF, aggregated over a domain, above the 90th percentile threshold for two consecutive days gave me quite decent results. The domain I am currently using is the same as that shown on the 26 July 2009 "Surge Classification" post. Some quick numbers before I shown the figures:
Average number of events in a given July or August: 5.6/month
This number is higher than what one should expect, but having a high number now is what I want. This is because I will also be agumenting this criteria with dewpoint thresholds which will bring this number down to are more reasonable number. And, actually....
Criteria: 2 days of IWVF >= 90th and 1 day of DWPT >= 90th
Average number of events in a given July or August: 1.1/month
This seems low, but plotting up the events for 1984 (the year that I have most been using during my studies so far) shows a very good agreement with what days I suspect are surge events. Below I will show those days (I based this on empirical evidence from Hovmoller diagrams which I am sure there are some on previous posts) and the plot for 1984 under the latest current surge criteria.
Suspected Event Days: ~38, ~50, ~70, ~78 and ~120.
Even though the monthly average over the 28 years is low it still gives a good representation of the surge events of this year. It could stand to reason that this low number is an artifact of the rather inactive monsoon years of the 1990's. I am basing my 'normal' per month surge events on Stensrud's word which is based on a time period of mid-70's through late 80's, typically.
Below is a composite of IWVF for a period 2 Days Before through 2 Days After an event. Map 3 corresponds to the event day, Map 1 and 2 correspond to 2 and 1 Days Before, respectively, and Map 4 and 5 correspond to 1 and 2 Days After, respectively.
From this data it seems clear that our IWVF is concentrated over regions adjacent, at least longitudionally, to the GoC. We can also see that as the surge happens parts of southern Nevada see very heightened concentrations relative to it's pre-surge values. Now, let's look at a composite of IWVF based on events of 1 Day of DWPT >= 90
th percentile.
Well, this is a very interesting and pretty cool figure. Again, Map 3 is the event, in this case
DWPT, day. Based on this figure, it appears to me that
IWVF is a precursor to
DWPT events. We see in the preceding days, i.e., Map 1 and 2, that the
IWVF field looks similar to the the post-surge event days, i.e., Map 4 and 5, from the
IWVF event figure above. This does make physical sense in that we would expect a large northward shift of winds before an increase of
dewpoint at the surface. I believe next I will use look at composites of
precipitable water. Since winds are not part of this variable, we will get a better idea of exactly how the atmospheric moisture is moving
pre- and post-surge events.
-
jamie
Some of the events do show a northward progression, e.g., Day 55. But, overall, most events appear to show up on all domains on the same day, e.g., Day 100 and 78. The later is the event that I have used most frequently for case studies because it was shown to be one of the most substantial northward surge events in our study domain. Here, it appears to all happen on the same day but with a lessening intensity further north. An event on roughly Day 28 seems to work the reverse to this in that it is stronger the further north you go. Since this event is in June it could possible be more synoptically influenced rather than monsoon related. These subtle differences might be a good way to further classify events as "Strong" or "Weak."
