Variation In Precipitation At Dulles International Airport (IAD) 1964-2017

Previously I wrote about the variation In precipitation At Washington Reagan National Airport (DCA) and Baltimore Washington International Airport (BWI) to see if there was a significant change over the years due to Climate Change, and now it’s time to look at Dulles International Airport.

The IAD dataset starts later than that for DCA and BWI, and starts from April 1, 1960. As with the DCA and BWI data, there appear to be a few days with missing data, as for some years there is only data for 364 days. The actual number of days per year of data was taken into account when computing means. Also, large amounts of data are not available for various dates 1960 through 1963, and of course there is no data for the remainder of 2018. So the plots are from 1964 through 2017.

Each of the graphs can be clicked to be viewed full size.

The first graph is the average daily precipitation. This is total amount of rain per year divided by the number of days in the year. This graph shows a slight increase, from 0.110 to 0.115 inches per day:

Next is a graph of the number of days per year with precipitation, followed by its inverse, the number of days without. The number of days with precipitation has increased from about 114 to 121, or 6 more days. The increases at DCA and BWI were 5 days, and they was over a larger number years. It currently rains about 115 days per year at DCA, and 118 days per year at BWI.

This leads to the next graph, the average rainfall on days when it actually rained. There is no apparent change at all – this graph is flat. So while it rains about 6 more days a year at IAD, the amount of rain, on days when it does rain, is the same. Actually if you squint, it looks like it might be very so slightly decreasing, but it is subtle. And probably going to be different once you add in another year’s worth of data.

Next, we can look at the standard deviation in rainfall amounts on days when it rains, standard deviation is amount of variation of a set of data values around the mean (average). My previous post has a link for more information about standard deviation, as well as the best standard deviation joke for math nerds.

The standard deviation rose from about 0.48 to 0.50, about half the increase seen at BWI. By comparison, the standard deviation was unchanged at DCA. If you torture the data enough, you can find possible reasons for the change. I notice there was a step change / increase in the standard deviation from 2005-2014, which seems to have since stopped. You can also spot a period in the 1990s when the standard deviation was low. I am sure 20 people can find 20 more possible reasons. It’s like a Rorschach test for meteorologists. Here’s the standard deviation plot:

Finally, we can look at the number of days that rained and had rainfall over two standard deviations (using the standard deviation for that year), treating this as significant or extreme rainfall events. The trend of this is completely flat:

To summarize the analysis of rainfall events at DCA, BWI, and IAD:

  • The number of days with rain is increasing, 5 or 6 more days per year, over the last half century or more, at all three airports.
  • The average annual measured rainfall is either decreasing (DCA) or increasing (BWI, IAD), depending on which station you look at.
  • The average amount of rain on days when it does rain is either decreasing (DCA), increasing (BWI) or exactly the same (IAD).
  • The number of days with significantly high rainfall, exceeding two standard deviations, is either very slightly increasing (BWI) or unchanged (DCA, IAD), depending on which airport you look at.
  • Variation In Precipitation At Baltimore Washington International Airport (BWI) 1937-2018

    Previously I wrote about the Variation In Precipitation At Washington Reagan National Airport (DCA) 1945-2018 to see if there was a significant change over the years due to Climate Change, and now it’s time to look at BWI Airport.

    The BWI dataset starts earlier than tat for DCA, and runs from July 1, 1937 to July 18, 2018. As with the DCA data, there appear to be a few days with missing data, as for some years there is only data for 364 days. The actual number of days per year of data was taken into account when computing means. Also, some data is not plotted for 1945 or 2018, namely the number of days with/without rain for those years, as a full year of data is not available.

    Each of the graphs can be clicked to be viewed full size.

    The first graph is the average daily precipitation. This is total amount of rain per year divided by the number of days. It shows an increase from roughly 0.108 to 0.121 inches per day from 1939 to 2018. That’s an increase of about 12%.

    Next is a graph of the number of days per year with precipitation, followed by its inverse, the number of days without. The number of days with precipitation has increased from about 113 to 118, or 5 more days. This is the same increase we saw at DCA, although it seems rain about 3 fewer days per year at DCA vs BWI (You picked the right airport to leave near, Walter):

    That leads to the next graph, the average rainfall on days when it actually rained. This shows an increase from about 0.35 inches to 0.37 inches. (The trend at DCA was actually negative, from 0.37 to 0.35 inches in fact. Exactly opposite. Hmm…):

    Next, we can look at the standard deviation in rainfall amounts on days when it rains, standard deviation is amount of variation of a set of data values around the mean (average). My previous post has a link for more information about standard deviation, as well as the best standard deviation joke for math nerds. The standard deviation in rainfall amounts at BWI is increasing. Looking at the graph, the increase seems to be mostly due to an increase during the 2005-2015 period, which has since ended. Is it a short term variation due to random processes, or part of a long term shift?

    Next we can look at the number of days that rained and had rainfall over two standard deviations (using the standard deviation for that year), treating this as significant or extreme rainfall events. The trend of this is completely flat:

    But, you might ask, isn’t the standard deviation slightly increasing each year? What if we use a fixed standard deviation value, like 0.50 which seems to be the mean value? Doing that, you do see an increase from about 9 to 11 days per year:

    Next up… Dulles Airport (which unfortunately has the shortest dataset of all three major airports)

    Variation In Precipitation At Washington Reagan National Airport (DCA) 1945-2018

    Recently the weather, at least precipitation wise, in the DCA area has been variable. It was very wet and rainy, then we had dry conditions for several weeks with essentially no rain. Now, it is very wet again. Are we seeing extreme (some may say historic) changes in the weather? Or are these just the usual variations?

    Are rainfall events are becoming less common, but more extreme? That is, it rains less often, but we get more rain when it does rain, rather than getting rainfall spread out evenly over time as it used to be. And maybe we’re getting more rain overall. Or possibly less rain, those two claims seems to depend on recent weather memory.

    To check this claim, data for Washington Reagan National Airport (DCA) was downloaded from the NOAA NCDC site and analyzed in several ways. You can download this data yourself, if you wish: https://www.ncdc.noaa.gov/cdo-web/

    The dataset runs from July 1, 1945 to July 18, 2018. There appear to be a few days with missing data, as for some years there is only data for 364 days. The actual number of days per year of data was taken into account when computing means. Also, some data is not plotted for 1945 or 2018, namely the number of days with/without rain for those years, as a full year of data is not available.

    Each of the graphs can be clicked to be viewed full size.

    First, is there a significant long term trend in the amount of precipitation at DCA? No, there does not seem to be (if you squint you may see a very small decline over time, the sign of this slope likely changes from year to year with normal variability in rainfall):
    Mean daily precipitation

    Second, are rain events becoming less common, but with higher rainfall totals from those events? That would mean we are seeing fewer days with rain, but more rain on those days. The following two graphs show the number of days without any rain, and the number of days with rain, defined as 0.01″ or more. First the number of days without rain. Which is not increasing, but actually decreasing:
    Days per year without precipitation

    And the number of days per year with rain, which of course is just the inverse of the previous graph. It’s raining 5 or 6 more days per year (sorry, Walter):
    Days per year with precipitation

    Third, what about the rainfall totals on days when it actually rains. Is that increasing, leading to more extreme rain events? No, it isn’t. It is actually decreasing, which makes sense considering the mean rainfall per year is essentially steady, and it is raining a few more days out of the year:
    Mean precipitation for days with precipitation

    OK, maybe it is mostly the same, but we’re getting a few more extreme rainfall events per year? Let’s look at the standard deviation of the rainfall amounts, again only for days when it actually rains.

    Standard deviation is amount of variation of a set of data values around the mean (average), there is an explanation here: https://en.wikipedia.org/wiki/Standard_deviation If you get this joke, you understand standard deviation: “Yo mama is so mean she has no standard deviation”

    Hmm, no, that is also steady:
    Standard deviation of daily precipitation

    One more thing, look at the number of days with very extreme rainfall. How about the number of days where the rainfall exceeded two standard deviations? That also seems to be flat:
    Days above two standard deviations

    Anything else we can check to see if precipitation is indeed getting more extreme in Washington DC? So far, it doesn’t seem to be.

    Note, the purpose of this analysis was not to try and discredit man made climate change aka AGW, which is certainly real. Only to see if claims of a noticeable effect on the precipitation patterns in the DC area can be confirmed, which does not seem to be the case.