You may have noticed that at times rainfall totals in the Almanac section or History page suddenly shift, usually downward. That’s because the automatic rain gauge runs a little wet. Oftentimes, I circle back and correct the amount with what I’ve measured in the manual gauge (~ 6′ away and 3′ lower).
Well, through the wonders of technology (a.k.a. Microsoft Excel), I was able to finally calculate the error and program a correction for it. Over the last 12″ of rainfall, it’s a pretty consistent 4-8% overage. It averages out to +6.5%. So I am subtracting that from realtime readings now. I’ll keep an eye on it to see if it changes over time.
For small events, it may not have any effect whatsoever. But in larger rain storms, it should allow for fewer corrections and, thus, less confusion.
The radar is down until approximately 3 PM on Wednesday July 1st for maintenance.
Be aware that starting Monday July 13th the radar will be down for approximately 2 weeks. Technicians will be swapping out the pedestal. See the NWS blog article for more information.
As noted in the about page, I regularly check our sensors to be sure they are within specifications. Over the last couple weeks I have been conducting tests.
Temperature was right on. That’s typical, as our sensor is pretty bulletproof. I have a platinum RTD digital thermometer that’s accurate to within 0.1° F (best to check on a cloudy, windy night to eliminate radiation as a factor). I also have a laboratory-grade aspirated psychrometer with a dry bulb thermometer that’s extremely accurate. It’s analog, so the biggest challenge is reading between the lines. But my tests show the station’s thermal sensor is within 0.5° during the day which is quite good. There’s also a backup sensor on site as well. At night or on cloudy/rainy days, the two sensors are normally within 0.2° F. During sunny days, height (7 ft vs 21 ft) and shielding differences (active vs passive ventilation) can frequently lead to 1° differences in either direction. 2-3° differences are not out of the question when it’s particularly calm and sunny.
Continue reading “Instrument Calibrations”
Earlier this week I removed the cloud icons that formed the main navigation links throughout the site. This design methodology had passed its shelf life. It also limited the font size of the embedded text. By deploying more generic buttons I was able to change its font type & size. This should make navigation clearer & easier to use.
In the weather history section, you can now click on each of the 24 hour graphs at the top for a much larger, more detailed version of the image. These images are not straight duplicates of the data in the smaller graphs and contain some custom “enhancements”. For instance, the detailed dew point and humidity graphs contain corrected & smoothed values (rolling 5 minute averages to prevent readings from jumping around too much when the air is unstable). The software that connects the weather station to the website cannot correct for sensor errors in real time. I do that immediately after the reading is taken. Then I send the corrected values to the site (current conditions, history tables). Unfortunately the smaller graphs are generated before I can access them. Note that the larger dew point and humidity graphs are still “initializing” which means that some of the data from yesterday is missing. This will correct itself by about 8 PM Saturday evening.
I also fixed a small bug whereby if you had selected metric units in the page settings of the City page (home page) AND you also selected unrounded temperatures you would only see rounded temperatures.
Usually I wait until May 1st, but the upcoming 7-day forecast shows temperatures almost entirely above freezing with daily highs in the 40s & 50s. Some rain is possible next week. This is not what I would describe as “wintry” by U.P. standards. Since the site only has two modes, I think it’s time for “summer” mode.
This means that wind chill will no longer display in the current conditions on the City page (home page). Rain measurement will take its place to the right of temperature. However, until April 30th, you can monitor hourly wind chill via the daily table on the Weather History page.
Keep in mind that for wind chill to be calculated the temperature must be under 46° and the average wind speed must be at least 3 mph. Otherwise, the wind chill is the same as the ambient temperature. In actuality, wind chill remains very close to (if not identical to) the air temperature until the latter drops into the 20s and/or there are very strong winds.
If we get another cold spell this spring, I will return the site to winter mode temporarily. Let’s hope that’s not necessary!
P.S. If it snows in the next 7 days, it’s officially my fault.
UPDATE 5/9/2020: The website is temporarily back in winter mode due to the recent snowfall and cold temperatures. I anticipate by Tuesday the 12th I can return it to summer mode for good (knock wood).