Raws Warm Bias Continued

A couple of weeks ago I presented a few results from my latest project - an attempt to adjust RAWS temperature data to remove the warm bias that occurs during strong sunshine.  The goal here is to make the RAWS temperature data more useful for climate monitoring; we want to know the spatial and temporal distribution of temperature variations across Alaska, but the RAWS measurements are heavily affected by this warm bias that varies depending on sunshine and - to a lesser extent - wind speed.

In the previous post I showed the results of a bias correction based on the hourly quantity of solar radiation, for 3 different RAWS sites that are located close to reliable FAA instruments (ASOS/AWOS).  Now let's look at the effect of wind speed, which is also measured by the RAWS platform.  The charts below show the residual differences between the RAWS and FAA temperatures after the solar adjustment has been applied, with hourly mean wind speed on the horizontal axis.  The red markers indicate the median difference for each wind speed value; note that wind speed is reported to the nearest whole number in mph.

At all 3 sites, increasing wind speeds cause the RAWS temperature to decrease relative to the ASOS temperature, which is what we expect; when a breeze is blowing, the thermometer is naturally aspirated and the airflow through the thermometer housing helps reduce the artificial warming from solar heating.  This means that the warm bias becomes less of a problem as the wind picks up, and therefore it also means that my solar adjustment is too great when the breeze is blowing: if I apply my solar adjustment without regard to wind speed, then my adjusted temperatures will be too low (as shown in the scatter plots).

The obvious next step is to model the wind speed effect in a similar manner to the solar effect, and I've done that using another analytical function to describe the relationship.  After optimizing the fit of the function for each site separately, the results look like this:

Happily we find that the average wind speed dependence has been largely removed, although of course this is not a perfect process; the temperatures still seem to be biased a bit high at low wind speeds at Eagle and Lake Minchumina.

So after all this we have a set of hourly adjusted temperatures for these RAWS sites, and we can now run a test to see whether the revised data show monthly or annual climate variations that are similar to those measured at the FAA sites.  Here we are interested not so much in the long-term average bias, which can always be removed by subtracting the long-term normals, but in the sign and magnitude of month-to-month and year-to-year changes.

Ideally we would find that such changes are very similar for each pair of sites; for example, when the adjusted Fairbanks RAWS data say that a month was 3°F warmer than normal, then we want to see that the ASOS data show the same anomaly.  If this is true, then the monthly temperature differences would remain constant over time - indeed the differences would be zero if the bias is fully removed - and then we could claim that the adjusted RAWS data provide a true estimate of the long-term temperature variations.

I'll start by showing results from Lake Minchumina, where the adjustment procedure seems to have paid off handsomely.  The first chart below shows the May, June, and July monthly means of daily high temperature before and after the RAWS adjustment; the FAA/AWOS temperature is also plotted in blue.  Note that these results are drawn only from the sample I used for the adjustment process - i.e. only "peak sunshine hours", so the high temperatures might be different from 24-hour values in some cases.  Clearly the adjusted RAWS numbers show very similar month-to-month and year-to-year changes to the AWOS data.  The unadjusted RAWS data also capture the major ups and downs, but notice that there's a trend in the differences: the unadjusted RAWS line is closer to the others in more recent years.

The chart below highlights the trend issue by showing the monthly mean differences of daily high temperature between the two sites, with the unadjusted differences indicated with solid lines and the adjusted differences shown with dashed lines.

The key thing to note here is that the adjusted differences don't change significantly over time - there is little trend and the monthly variance is much smaller than for the unadjusted data.  This means that, as we saw above, the adjusted RAWS temperatures essentially move in lockstep with the AWOS temperatures.  This is in contrast to the unadjusted RAWS data, which show a remarkable trend: the RAWS warm bias has diminished considerably in the past few years.  We might be tempted to speculate about instrumentation changes as a cause for this, but the fact that the bias correction eliminates the trend suggests that solar radiation has been reduced significantly in recent years.  Obviously I'll have to confirm whether that is the case; it would be an interesting result by itself.

In conclusion, the removal of the RAWS warm bias at Lake Minchumina appears to work very well as a means to improve the quality of the data for climate monitoring.  Unfortunately, the monthly mean temperature results are not as encouraging for Fairbanks and Eagle - see below.  The long-term average bias has been removed, but the monthly temperature differences are not significantly less variable than for the unadjusted RAWS data.

The disappointing results at Fairbanks could be related to the fact that the RAWS and airport ASOS sites are over 12km apart, in contrast to Lake Minchumina where the two sites are only a few hundred meters apart.  I looked into using Fairbanks' Fort Wainwright ASOS instead of the airport, but the Fort Wainwright historical data are not as complete.

In Eagle the problem could simply be that the solar warm bias is smaller, as shown in the first post, so there's less opportunity to improve the RAWS data.

Finally, as a measure of the degree of improvement, here are (1) the correlations of the monthly mean temperature anomalies before and after adjustment, and (2) standard deviation of the monthly mean temperature differences before and after adjustment.  The higher the correlation and the smaller the standard deviation, the better.

Site	Correlation: before (after)	Standard deviation (°F): before (after)
Lake Minchumina	0.93 (0.99)	1.27 (0.43)
Fairbanks	0.91 (0.92)	1.56 (1.37)
Eagle	0.98 (0.98)	0.79 (0.70)

Modest Fire Season So Far

So far the Alaska fire season has been fairly subdued although certainly not inactive.  According to today's report from the Interagency Coordination Center, the statewide fire acreage has now exceeded 300,000 acres, which is just over half the 1996-2016 median for the date.  The acreage is increasing by about 25,000 acres a day at the moment, but this is mild in comparison to what happened two years ago, when about 2 million acres burned in the first 10 days of July.  See here for a blog post from the time.

The only two large fires in the state are burning within the Arctic National Wildlife Refuge, with one of them having crossed over from Canada (see more details here).  Widespread rain has fallen in the central and southern interior recently, but the northeast of the state has been hot, dry, and windy; for example, the Fort Yukon SNOTEL site has reported high temperatures in the 70s and 80s for 3 weeks now, with only 0.3" of rain in that time.  Similarly, Old Crow in the northern Yukon Territory has seen highs above 80°F on 4 of the last 5 days including today.

Satellite Perspective on Arctic Lightning

In yesterday's post on lightning over the Beaufort Sea I somehow neglected to look at the satellite imagery to confirm the presence of deep convective storms in the region where lightning was reported.  I realized this oversight when Rick Thoman sent me an image last night - thanks Rick!  After digging around on the UAF Geographic Information Network of Alaska (GINA) website, I found several images that clearly show the evolution of the thunderstorms.

Here's a particularly nice photo showing a curved line of storms in approximately the center of the image; the shadow of the deep cumulonimbus clouds is evident to the north of the line (the sun's position is in the southeast at the time of this shot).  For reference, Point Barrow is visible in the center left of the image, and Banks Island is towards the upper right.

Here's a zoomed-in photo.

In the image sequence below I've paired up a selection of satellite views with the lightning reports from approximately the same time, running from about 2am to 1pm AKST on Monday; the lightning strikes clearly correspond to the brightest (deepest and therefore most reflective) storm clouds.

Arctic Lightning

The other day a reader pointed out that the Alaska Lightning Detection Network (ALDN) showed a notable cluster of lightning strikes on July 3rd over the Arctic Ocean north of the Mackenzie Delta.  Here's the 24-hour map of reported strikes:

There was also lightning in the same general area the day before and the day after, if the sensors are to be believed.

The 500mb charts show that an upper-level low pressure system drifted up from the southeast and lingered near the Mackenzie Delta for several days, so this may have provided the instability necessary for thunderstorm activity; the favorable meteorological setting increases the likelihood that thunderstorms really occurred and the reports were not just an artifact in the detection system.  Here are the 500mb maps at 24-hour intervals from Saturday morning (July 1) through yesterday morning (July 5).

To get a sense of how unusual it is to have this much lightning so far north, I examined the ALDN data since 2012.  The ALDN historical data goes much farther back, but in 2012 the ALDN switched to a new detection system with greater sensitivity and greater range (see this 2012 news article); so the data since 2012 are not comparable to the data from earlier years.

Looking at lightning strikes north of 70°N, it does appear that this week's event is unprecedented in the 2012-present history of the ALDN data.  On July 3 alone, 1521 strikes were reported north of 70°N in the ALDN domain, and the previous record was 1011 on June 24, 2012.  Here's the map from that day in 2012; note that the blue markers show strikes recorded by the old detection network.

It's very interesting to see that the same portion of the Arctic Ocean was affected in 2012, but a long-term study would be needed to see if this region is climatologically favored for electrical storms.

Here's a chart of the monthly strike numbers north of 70°N for May through August of 2012-present.  With only 5 days of data so far, this month already has more strikes than June 2012, and no other month comes close to this level of activity.

Is it a coincidence that 2012 saw high lightning activity and also produced the most recent record low sea ice extent in September?  Perhaps not; the ice extent maps below show that 2012 and this year both had widespread open water north of the Mackenzie Delta on July 1.  One would expect open water to provide a more favorable storm environment than an ice-covered ocean, via increased evaporation and greater low-level moisture.

Some of the intervening years were not too different (see below), but nevertheless the similarity between 2012 and 2017 is rather striking.

RAWS Warm Bias

Regular readers will know that Alaska has a reasonably dense network of realtime temperature observations these days from remote automated weather stations (RAWS) that have been deployed by various agencies.  Data from 92 of these sites is archived in NOAA's GHCN data set, but the Western Regional Climate Center lists more than twice as many stations in Alaska on their website.

As someone who is interested in climate monitoring for Alaska, I've been thinking about whether it's possible to make the RAWS temperature data more useful for climate monitoring purposes.  For example, back in April we noted that NOAA's climate division analysis for March temperatures in the North Slope region seemed quite unrealistic; the reasons for this are not clear, but it's tempting to think that the modern wealth of RAWS data could help provide a better picture.

However, quality is a perennial issue with RAWS temperature data.  One of the most significant problems is one that we've mentioned many times here; RAWS thermometers are not artificially ventilated, so they nearly always read too high when the sun is shining strongly and winds are light.  The difference can be on the order of 10°F, which renders the data much less useful for climate monitoring.  This kind of issue can really degrade the analysis and conclusions if one isn't careful; here's a paper showing major impacts from a different kind of bias in SNOTEL temperature data.

Oyler et al., 2015: Artificial amplification of warming trends across the mountains of the western United States

I decided to take a closer look at the RAWS solar warming effect by examining data from RAWS sites that are located close to higher-quality FAA instruments.  So far I've identified just 3 locations where the RAWS and ASOS/AWOS thermometers are in fairly close proximity and at similar elevation: Fairbanks, Eagle, and Lake Minchumina.  Here are scatter charts of daytime hourly temperatures from the past 8 years, for a calendar period of about 45 days either side of the summer solstice; this is the season of peak solar heating (although cloudiness increases as summer advances).  I've also restricted the analysis to times of the day (about a 10-hour period) in which the climatological normal solar radiation is above a certain threshold - what I'm calling peak sunshine hours.

As we would expect there's a very good correlation between hourly temperatures at all three sites, but in the case of Fairbanks there's a distinct indication that the RAWS temperature is more often significantly higher than the ASOS temperature when the weather is warmer.  We would expect to see this if strong sunshine creates a warm bias at the RAWS thermometer, because hours with strong sunshine will tend to be warmer anyway (all else being equal).  There's a hint of the same relationship at Lake Minchumina, but not at Eagle.

To highlight the dependence on solar radiation, I plotted up the temperature difference at each pair of sites against the hourly total solar radiation as reported by the RAWS instruments - see below.  Now the relationship is quite striking and there's no question at all that strong sunshine tends to produce a substantial warm bias in the RAWS temperature measurements.  The charts for Fairbanks and Lake Minchumina are quite similar, but the bias is much less evident at Eagle.

In an attempt to describe the relationship mathematically, I came up with an analytical expression for the temperature bias and then fitted the coefficients of the expression using an optimization routine for each site separately.  Here's what the curve looks like; it seems to do a reasonable job of describing the effect of solar radiation.

Finally, here are the residual temperature differences after the RAWS temperature data are adjusted to remove the solar-caused bias.  As we would hope, the systematic effect of the solar radiation has been removed so that the adjusted temperature differences are independent of solar radiation.  Note that the odd curves that are evident within the distribution of data points are just an artifact of starting with whole-number temperature data.

The next step in the analysis will be to look at the impact of wind speed, because breezy conditions tend to reduce the warm bias that occurs when it's sunny; look for some more scatter plots in a subsequent post.

But for now, this is an encouraging start - in particular, the similarity of the solar radiation impacts at Fairbanks and Lake Minchumina suggests that it may be possible to apply a universal correction to all temperature data originating from RAWS instruments that are similar to these two sites.  I'm not sure what's going on at Eagle - the relative absence of a solar warm bias is mysterious - but perhaps there's a difference in instrumentation.

And finally, if anyone knows of any other RAWS installations that are close to FAA sites, let me know - it would be great to add more sites to the analysis.