Analysis of Biases in the Official Fallbrook Temperature

Introduction
Analysis
Summary


Introduction

Beginning sometime in 1999, it became clear that something was funny about the official Fallbrook high temperatures, especially in comparison to the Vista temperatures. Anyone watching the evening weather report that showed the official high temperatures across the county noticed that the Fallbrook high temperature seemed much too cool to be believed. I was not alone in noticing this discrepancy. Larry Sundram also noticed this change and wrote me. Undoubtedly, many others also noticed it.

On normal warm days (without the influence of storms), the Vista, Escondido and Fallbrook high temperatures are within a few degrees of each other. All three are normally significantly warmer than the coastal temperatures of Oceanside and Carlsbad. Since the evening weather report plots the high temperatures on a map of the county, it is easy to see these groupings of high temperatures.

Prior to 1999, these groupings followed the usual pattern. But in 1999 a change became apparent. The Fallbrook temperature no longer agreed with those of Vista and Escondido to within a few degrees, and was cooler by ~5° on many occasions. There were even some non-winter days in which the reported Fallbrook temperature was colder than Oceanside or Carlsbad, a very unlikely situation!

The following table gives some examples of the official high temperatures, randomly selected as the last few days for which I still had old newspapers around with reported temperatures for all these cities:

CityHigh Temperature (° F)
6/4/006/13/006/14/00
Oceanside686873
Carlsbad727169
 
Fallbrook827979
Rainbow868383
Vista868380
Escondido888684

(The North County Times reported a Vista high temperature of 90° on 6/14/00, but the true high temperature that day was 80°. In general, one must take considerable care in using the temperatures reported in the North County Times, since they often make mistakes in reporting the previous day's temperatures and rainfall amounts. I wrote them several times about errors on their page, but gave up due to the frequency of such events and after never receiving any reply.)

Note that the reported Fallbrook high temperature is 1-6° cooler than Rainbow, Vista and Escondido on all three days, a pattern that has persisted for the last year or so. This is quite different from previous years. For example, in 1997 I found that Fallbrook usually reported a high temperature that was hotter than that of Vista by 2-5°. Thus something has changed, making the Fallbrook - Vista temperatures different by ~5°.

The analysis below shows that in fact two changes have occurred:

This is not a result of global warming or any true changes in the weather! It is instead surely due to changes in equipment or the calibration of the equipment at both stations.

Analysis

Gaylord Moxon and I also record the daily temperatures at our Fallbrook homes. I have analyzed the biases compared to the official Fallbrook high temperature previously, so it is a simple matter to repeat that analysis to see if the biases have changed. With three or more thermometers, it is possible to definitively identify which thermometers have changed.

A plot of the median bias of my high temperature vs. the official Fallbrook high temperature clearly shows that in March 1999 the bias changed by ~2-3°. (The sharp dip in July 1998 is caused by a large amount of missing data, making the numbers scatter more.) Prior to March 1999, the bias was 2-4°; after March 1999 the bias was 5.5-7.5°, an increase of 3.5°. A similar plot comparing my high temperature vs. the high temperature for another thermometer that I also record in my yard shows no change during this period. Hence I conclude using my own thermometers that the reported Fallbrook high temperature got cooler by ~3.5° around March 1999.

Gaylord Moxon's high temperatures confirm this result, although it is a bit more complicated due to the yearly variation in the bias on Mox's high temperatures and because Mox's environment has changed slightly with time. The changes in Mox's environment are that his cover on the porch lattice over his temperature sensor was replaced in early 1998, changing it from plastic to wood, and some nearby large eucalyptus trees were removed in October 1999.

The blue line plots the median difference between Mox's high temperature and the official high temperature. It shows a yearly variation very similar to that found previously. The pink line in the plot is the difference in the bias between the current date and a year earlier. Thus the pink line attempts to remove the yearly variation, if there were no changes, and shows any change in the bias with time.

Before March 1999 the bias is consistent with zero. After March 1999 the bias is 1.7-3.8° until the eucalyptus trees were cut down in October 1999, making Mox's high temperatures significantly higher. These data thus also imply that the reported official Fallbrook high temperature has gotten cooler by about 2.8°.

Summary

The reported Fallbrook high temperatures are now ~3° cooler than prior to March 1999. This is not related to any change in the weather, and was derived solely by comparing four thermometers making measurements on the same day over a yearly period. Of the four thermometers, only one - the one recording the official Fallbrook high temperature - changed its calibration in March 1999.

Mox and I visited the location of the official Fallbrook thermometer at the downtown fire station at 315 E. Ivy on 6/13/00, and found that the temperature is now being taken by a wireless modern electronic sensor located some 15' or so off the ground. This is a new sensor as of sometime later than September 1998. We are betting that this new sensor went into service in March 1999.

Apparently the official Fallbrook thermometer used to be computerized prior to ~1996, but it broke around then and was not replaced until 1999. Temperatures between 1996 and 1999 were manually recorded using the standard white box ~4' off the ground.

Hence it seems likely that the change in the reported Fallbrook high temperatures was due to the replacement of the official thermometer in March 1999. I cannot determine for sure whether the current temperatures are too cool, or the previous temperatures were too warm. There is some evidence from other analysis to be reported later that the current temperatures are too cool.

This doesn't explain the apparent change of ~5° in comparison with the Vista high temperatures. (The number of 5° is based on far less data than the bias of 3° found here, and hence is a less accurate number.) Interestingly, detective work by Larry Sundram revealed that the Vista weather station has changed twice recently. During 1998, it was moved from Fire Station #3 (located on Old Taylor Road near its intersection with East Vista Way) while the firehouse was being remodeled. It was then moved back in late 1999. However when it was moved back its surroundings had changed. It now gets a fair amount of reflected heat from a building constructed of different material than previously and from a concrete parking lot.

Interestingly, on 6/6/00 the National Weather Service told Vista that their reported temperatures seem to be 3-5° degrees higher than they should be. As a result the fire station is going to move the temperature gauge about 30'. I suspect that some of that 3-5° bias comes from the new bias in the Fallbrook temperatures. Stay tuned until there is enough data to sort this out.


Go to Fallbrook Weather


Copyright © 2000 by Tom Chester.
Permission is freely granted to reproduce any or all of this page as long as credit is given to me at this source:
http://sd.znet.com/~schester/fallbrook/weather/analysis/official_temperature_biases.html
Comments and feedback: Tom Chester
Last update: 15 June 2000.