Urban Heat Island and the Arizona Monsoon

Over the past week or two, Arizonans began their annual Monsoon speculation. For the next few months, Arizonans will devote considerable casual conversation to the weather – when it will rain, why it won’t rain, where it will rain, and where it did rain. The conversation rarely leads anywhere, but provides welcome relief from tired topics such as the baseball and politics.

For Phoenix area residents, the conversation focuses on one observation – that the Monsoon rains were more frequent and intense “back in the good old days” when the city was smaller. Most long-term residents have similar observations. I remember residents discussing this “diminishing Monsoon” phenomenon for most of my 27 years as an Arizona resident. A multitude of hypotheses accompany this observation, such as extensive concrete and asphalt, air pollution, and urban heat islands blowing the storms away. Do these hypotheses have merit? Is the “diminishing Monsoon” observation of long-time residents validated by rainfall measurements?

Thankfully, the internet makes rainfall data acquisition a simple task . The National Weather Service (NWS) and other agencies record this information and make it available to the public via the internet. Data is easily downloaded and examined using software most people have on their home computers. Rather than speculating or supposing, ordinary people can take raw data and test Monsoon hypotheses for themselves.

I obtained Monsoon rainfall data for the Phoenix metro area from the National Weather Service HERE. This rainfall data comes from the official gauge (now located at Sky Harbor airport) near the center of the city. This gauge should record any effect of urbanization upon rainfall. Acres of concrete and asphalt, air pollution, and sparse vegetation surround the airport gauge. The Maricopa County Flood Control District also maintains and makes available on their website data from a network of neighborhood gauges. I chose the NWS data because I felt it sufficient for a simple analysis.

The NWS Phoenix Monsoon data covers years 1896 to 2009, and from June 15 to September 30, the currently defined “Monsoon Season”. The NWS also breaks down rainfall by month. Data is also available for other southwest US cities, and that may be useful for comparative studies. For now, I am only looking at Phoenix data.

Plotting the total Monsoon season rainfall in Microsoft Excel shows that monsoon rainfall is extremely variable from year to year. Looking at the graph shows no obvious trends of either increasing or decreasing rainfall. A linear trendline through all the data shows a very slight decreasing trend with a slope of -0.0069 inches per year. This trendline is partially skewed
because the first year (1896) was much wetter than average, and the last year (2009) was much drier than average. The important observation is that the negative slope is small, and the overall graph does not show any obvious trends or inflections as the city has grown.

Since I am testing people’s personal observations, it is useful to examine the period of record for only the years people remember. I chose 1960 to present, figuring the “good old days” for most people were the 1960 and 1970s. Phoenix also grew rapidly during that interval, so an examination of data from 1960 to present should capture the decreasing rainfall trend described by local residents. (A comparison of Monsoon rainfall versus urban land area or population is certainly possible. See the link at the end of this post).

Again, the chart reveals high annual variability in Monsoon rainfall and does not show any obvious trends or inflections. A linear trendline through the data has a very slight negative slope of -0.0039 inches per year. Note the shallower slope than from 1896-2009. Considering that 2009 was significantly drier than average, the trendline may again be influenced by anomalously wet or dry years near the ends of the period examined.

To illustrate the influence of data points near the ends of the data set, I also plotted 1960-2008. A trendline shows a positive slope of 0.0002 inches per year. A one year adjustment in the data range reverses the rainfall trend from negative to positive. This highlights the weakness of the slightly decreasing trends observed in the 1896-2009 and 1960-2009 trendlines. Also noted is the importance of choosing an appropriate period of record when examining data. Statistics are easily manipulated if the examiner chooses a data set to only validate a particular hypothesis. Reasonable conclusions consider the overall strength of a trend, and the strength of any correlation to other data. In the case of annual Phoenix Monsoon rainfall, no strong trend or correlations can be made from these data. Other studies may reveal different trends depending on what data set is used, in comparison to population or in examination of spatial rainfall distribution. Such studies could reveal different trends, but those trends are not likely strong enough to be evident to the casual resident.

Some may point to the dry years from 2000-2005 of evidence of diminishing Monsoon rainfall with urban expansion. Drawing a conclusion for those six years is tenuous because:
1) The trend did not continue during subsequent years;
2) Similar dry periods occurred in the early 1920s and early 1930s;
3) Local residents were discussing diminishing monsoon rainfall long before 2000.

The basic rainfall data examined here do not show a trend that validates observations of long-time residents that Monsoon rainfall has decreased with expansion of the Phoenix metro area. .

If data studies do not corroborate the almost universal observation of decreasing rainfall with urban expansion, what causes the observation? I hypothesize it has to do with a “good old days” phenomenon where humans tend to remember things differently than actually occurred. Several reasons can contribute to this inaccurate recollection:

1) People tend to remember the most significant events from the past, and forget ordinary events. Thunderstorms and wet summers are remembered, whereas average or dry summers are forgotten.
2) Most Phoenix residents are transplants. For those outsiders, the Monsoon is a strange weather phenomenon in that months of bone dry weather is interrupted by a tropical influx. These newcomers may remember their first few Monsoon seasons to be more significant than they actually were. After a few years, newcomers adjust to Arizona weather and the summer weather phenomenon becomes ordinary. Suddenly, the Monsoon does not seem as wet as it used to be.

Finally, I am aware that studies have been done relating urban heat islands to summer thunderstorms. Some studies focused on the Phoenix metro area specifically. The goal of my blog is to illustrate that basic scientific curiosity and investigation is available to the average person. In other words, value exists in having people “see for themselves” instead of attempting to read a study with language and term they don’t understand, or relying on interpretation of those studies by popular media. For those that wish more information or results of a more rigorous study, I suggest starting with this online article from the National Oceanic and Atmospheric Administration:
NOAA Monsoon Study

This entry was posted in investigations, weather and tagged , , , , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s