Sitka Precipitation Records

(follow-up with corrected precipitation totals)

Recently Sitka’s precipitation totals have been in the news as part of the on-going issues related to relatively water low levels in Green Lake and Blue Lake, where the vast majority of our electricity comes from in this community. As a result, the electrical department has been utilizing diesel generators, a much more costly alternative to hydro. The most recent news item on Raven Radio included a link to a NOAA page with a graphic showing the rainfall to date compared to normal.

(plot from NOAA website generated 5 December 2011)

The plot shows that up through the beginning of December Sitka has seen about 58 inches of rain, roughly 20 inches below normal. This is all well and good, and would certainly seem to explain the low lake levels. I didn’t think anything more about it until a fellow natural history enthusiast and friend reminded me that not that long I had mentioned looking up the precipitation records for this year, and they were appearing to be pretty typical in terms of yearly totals. Of course, as is often the case, I ended up spending quite a bit more time looking into the question than I might have initially intended (and, as it’s my current policy to write on here about whatever nature-related things I happen to be thinking about – today it’s the rain).

Sitka proper has 4 stations that have recorded weather data over the years (there are others on Baranof Island as well). The longest running one was at Geodetic Way – that record started in the 1800s. At some point along the way, a station was started at the airport, and in more recent decades it has been the primary source of data. For a period of time in the 2000s the Fire Station had a weather station, and the Waste Water Treatment (WWT) plant has been recording data as well since 2006 or 2007. One of the things that was interesting to me, is comparing the rainfall records between the WWT Plant and the Airport, locations which are pretty close to each other, the amounts are similar, but with noticeable differences. Topography can make a huge difference in rainfall totals, but I don’t think that’s much of an issue in this case. I suspect it’s more due to the randomness of rain shower tracks, and some error due to differences in measuring, perhaps (the airport is an automatic station, I’m not sure how the WWT plant one is done). I didn’t check the other stations, as neither were recording data this year, and I was more interested in figuring out the discrepancy in totals for this year.

The data I had looked at previously via Weather Underground’s weather almanac for Sitka Airport, and indeed shows that totals are pushing toward 80 inches for the year – a total that is about as close to average as you could want. Looking at the monthly totals from the Weather Underground record, things seemed to square with my (admittedly vague) recollection of what happened. I decided to see if I could figure out where the NOAA plots were coming from.

By going to the Alaska Climate Database and selecting the appropriate options, I able to generate the following table that had numbers consistent with the plot (I never did figure out how to get the plot generated):

On the other hand, using the NOAA Online Weather Data (NOWData) option and choosing Sitka, monthly avgs/totals of precipitation for the current year I got the following table:

These numbers were pretty much in line with what I saw on Weather Underground.

Checking the months individually, it’s not hard to see that many months are identical, but June, August, October, November, and December are all different, with the Weather Underground/NOWData always greater than the Alaska Climate Database data. I checked the NOWData for November with the Alaska Climate Database, and the latter had several missing days which appeared to account for the differences. I did not see a way to access the August NOWData – but comparing the August 2011 data on Weather Underground with the Alaska Climate daily data, it seemed like they were two different months. Since I remember this August being much wetter than 2.93 inches, I suspect the data in the climate record was mislabeled in some way – perhaps a different station, or a different month or year.

It’s worth noting that the Alaska Climate Database page has a disclaimer which says: “Please note this information is preliminary and subject to revision. Official and certified climate data can be accessed at the National Climatic Data Center.”

In the end, my guess is that this year’s precipitation totals are pretty close to average. This year’s low water is probably due to 2010’s low precipitation (Weather Underground and the Alaska Climate Database have slightly different numbers, but both seem to agree that totals for the year were 20 or so inches below average), perhaps coupled with a cool summer and less than typical snow melt (see below). This may explain in part why many people (myself included) have been a little bit puzzled when hearing that reservoir levels are down because it’s been a low water year this year – it really hasn’t seemed particularly dry to me, and I’m thinking the records probably bear that out. Now that I’ve looked into things further, I imagine we were running a significant deficit after last year, and so needed extra water this year to make up the difference. (Which makes some sense to me – I’m thinking now it might be interesting to cook up a simple numerical model to see when the effect of dry periods hits – the year it’s dry, or the year after?) It seems like I’ve also heard that the water year is different than the calendar year – but I don’t recall what the dates are for the water year (so it’s possible when the electrical department says “this year” they don’t mean quite the same thing that I do).

Of course there are at least a couple of other potential confounding issues in all of this. One is the assumption that the precipitation at the airport correlates well with the precipitation in the Blue Lake and Green Lake watersheds. This seems like a reasonable assumption over the course of the year (though perhaps not in any particular storm) – however it also seems plausible to me that there could be a ‘normal’ year at the airport that was below average in terms of mountain precipitation (or vice versa) just due to random variation. A related, but perhaps even more likely issue is an effect due to snow pack. The amount of snow deposited in one year may not correlate strongly with the amount that melts. This past summer wasn’t particularly warm, so perhaps some of the low lake levels can be attributed to more snow remaining on the slopes. I suspect the data necessary to investigate these issues probably hasn’t been recorded.

I found this question interesting to investigate. It’s caused me to reflect on what we mean by weather records for a city – of course I know that temperatures and precipitation vary from place to place – even on a small scale, but despite knowing this, I still find it easy to forget that when I am considering whether it’s an ‘average’ year in town. The numbers I use to evaluate that question are really based on measurements at one point. It’s certainly not unreasonable to think that point reflects conditions across a broader area, but there is no guarantee that the broader area is particularly large, nor that any differences are consistent from season to season or year to year. More generally, it’s probably these relatively small scale differences that drive some of the variation in plant and lichen occurrence (and perhaps vice versa, though that seems less likely to be the case here than in more arid climates). It’s these sorts of patterns I think will be interesting to try to investigate and more fully understand in the coming years.