This week we welcome long-time friend of GLOBE, Dr. Peggy LeMone, Chief Scientist for the GLOBE Program from 2003-2009, as our guest blogger. Dr. LeMone is currently working in the field of weather and cloud formation at the National Center for Atmospheric Research (NCAR).

Originally posted at http://spark.ucar.edu/blog/measuring-rainfall on September 23, 2013.

Dr. Peggy LeMone is an NCAR Senior Scientist who studies weather and cloud formation. For more information about her research, visit Peggy’s home page.

A guest post by NCAR scientist Peggy LeMone

The Boulder, Colorado area received huge amounts of rain in mid-September.  You also learned that rainfall amounts vary a lot. Which brings us to the questions – How do you measure rain?  And how accurate are the measurements?  Even though I have done weather research for many years, during this storm I was reminded how hard it is to measure rain accurately.

This is the story of my attempts to measure rain during the storm. It’s also about the many possible sources of error when making rain measurements – from old rain gauges to growing trees and even, possibly, inquisitive raccoons.

By Monday morning (September 16), I had measured over 16 inches, or 405 millimeters (mm), in our backyard rain gauge from the storm which began September 10.  The gauge is the same type the National Weather Service uses. It has a funnel that deposits rain into an inner tube with a smaller diameter (like this one), but bigger. The inner tube’s diameter is just small enough to make the depth of rain ten times what it would be in a gauge without the tube and funnel.  Thus, each inch in the tube is equivalent to 0.1 inches (a tenth of an inch) of rainfall.  This is equivalent to how the GLOBE rain gauge measures rain: the inner tube acts like a 10x magnifying glass for the area of the rain gauge.  This makes it easier to read accurately!

My gauge is old. I inherited it from a weather-observing neighbor who moved away.  The funnel and inner tube doesn’t quite fit, so, I leave the gauge open and then pour the rain into the inner tube using the funnel.

On the morning of September 12th, the gauge was so full and heavy, with over seven inches (178 mm) of rain that I decided to stick a meter stick in the gauge to measure the rain amount, and save pouring into the inner tube for the end of the storm.  The gauge tilts slightly, so I took a measurement on the uptilt side and the downtilt side and calculated an average.   That evening I found that the bottom of the gauge sagged in the middle, leading to an even deeper measurement than the downtilt side.  With these flaws, the lack of the ten-to-one exaggeration of depth, and some measurements being taken in the dark with a flashlight, my data were only approximate. I recorded measurements to within the nearest quarter inch (see the graph below).

Were my measurements accurate? On Friday morning, September 13, I took measurement using a more accurate method to compare with my estimates.  After bailing out five full tubes of rain, I poured the remaining water through the funnel into the tube to a depth of 13.5 inches (343 mm), spilling a little bit during this process.  The result was 0.38 inches (9.5 mm) more than my rough estimate from the night before – a storm total of 14.52 inches (369 mm) up to this time. On the graph, this is marked as 1. (The lower shows the uncorrected values.)

But the rain hadn’t stopped.  I awoke on the morning of September 15^th and heard reports that up to 2 inches (51 mm) of rain fell overnight. I went outside to check our gauge – only to see that it had been knocked over (probably by raccoons).  Fortunately, I have a second rain gauge in my backyard – a plastic gauge that registered about 0.25 inches (6 mm). I added a conservative 0.2 inches (5 mm), since this gauge was under trees (marked as 2 on the graph).

The final number:  16.37 inches (416 mm) of rain, more or less.

The exposure of the rain gauge is undoubtedly the greatest source of error.  According to the National Weather Service and CoCoRAHS (both of which use citizen volunteers to measure rainfall), “exposure” of the rain gauge is important. Rain may be blocked by nearby obstacles causing the number to be lower than it should. Or, rain may be blown into or away from the gauge by wind gusts.  The recommendation is that the gauge be about twice the distance from the height of the nearest obstacles, but still sheltered from the wind.

The gauge was certainly sheltered from the wind.  It is located about 10 feet (~2 meters) south of the house, which is about 15 feet (5 meters) high, and to the west of a fence and small trees as well as the tree in the photograph.   There is a much smaller tree to the southwest.

All the obstacles suggest that some rain could have been blocked from reaching the gauge, which would imply that the rainfall total is too small.  On the other hand, some rain might have been running down the branch in the picture. (In fact, because of the large amount, I thought this might be the main effect before doing some research on exposure)

It is also recommended that the gauge be level, which it wasn’t.  I’m not too worried about this, since it was nearly vertical.

The conclusion?  There was a lot of rain.  It could have been an inch (25 mm) more or less than my measurement. Acknowledging this is called reporting error. It doesn’t mean that the measurements are wrong, it just gives an idea of how accurate they are. My total was not the largest; there were at least two other measurements near 18 inches (457 mm).

Now that I’ve described all that can go wrong measuring rainfall, let me add that, putting a rain gauge in the right place, and taking an accurate rainfall measurement is fairly easy. If you have a perfect cylinder, such as a GLOBE rain gauge, simply stick a ruler in and read the depth (make sure to correct for any offset of the “zero” line and correct for this offset; and see if the ruler pushes the water level up very much).

If you don’t have a rain gauge but have a bucket (or glass) with sides that aren’t straight up and down, you’ll need to do a little math to figure it out. Here’s what you’ll need to do:

1. Measure the diameter of the bucket at the level of the rain.  Subtract out twice the thickness of the walls.
2. Measure the diameter of the bucket at the bottom in the same way.
3. Calculate the average of the two diameters.
4. Divide by two to find the average radius.
5. Find the average volume of rain = Depth x radius x radius x 3.14.
6. Find the area at the top of the bucket (this is the area over which the rain is collected).
   1. Measure the diameter
   2. Divide the diameter by 2 to get the radius
   3. Area = radius x radius x 3.14 (remember that Area = pi x radius²)
7. Divide the rainfall volume by this area to get the rainfall.

It would be an interesting activity to put several buckets (or rain gauges) in different places in a field, your back yard, or your schoolyard to see how much the measurements vary within the area. Soup cans, though not perfect, would work pretty well for the activity, especially if they’re the same size.  I might try this during the next rainstorm.  (I hope not too soon!)

Does your school collect precipitation data? Have you had an extreme weather event that you were able to record? Let us know by adding a comment!

Why is zero so important?  Mathematically, zero is often used in reference to calculations and measurements.  In the scientific world, we often consider mathematics as the language of science; therefore, in science it also refers to a calculation or a measurement.  In many cases, knowing the absence of something is sometimes just as important as knowing the presence of something, such as migrating animals or even rainfall.

Let’s look at two real world examples that indicate how important zero is to Earth science:

First, let’s look at the Marshall Islands: The Marshall Islands, a collection of coral atolls and low-lying islands in the North Pacific, are undergoing a unique climate shift.  These islands, which comprise about 181 square kilometers of land, are experiencing extreme drought and extreme flooding simultaneously.  The following two pictures were taken only a month apart, but on two different islands, depicting two very different conditions.

Flooding in Majuro, Marshall Islands, June 2013. Photo taken by Anole Valdez, 2013.

Drought in the northern islands from May 2013. Photo from UN.org.

To put these two extreme events in perspective, observe the following map of the Marshall Islands.

Map of the Marshall Islands, from cia.gov

If only one atmospheric station reported data in a situation like this, then one could assume that the entire country was experiencing the same weather; Additionally, if atmospheric observers didn’t value the submission of zero rainfall and only reported data when there was rainfall, those not directly connected to these islands might think that it always rains and on certain days, observers fail to report their data.  That is obviously not the case, as can be seen in the two photos.  It’s important to collect and submit all observations to understand the entire scenario.

Second, let’s look at the Great Lakes region of New York State: In September of 1996, flooding occurred north of Buffalo, New York due to a lake effect rain event.  A lake effect rain occurs when the air temperature is much colder than the water temperature of a nearby lake.  As the cold air passes over the warm lake, some of the lake’s water evaporates and precipitates out (as either rain or snow, depending on how cold the air temperature is) over areas downwind of the lake.  In this specific example from 1996, Lake Erie observed water surface temperatures of 22.8°C, while the air temperature overnight was around 8°C.  As you can see by the radar image below, there was only a small area of land that experienced this extreme weather event (within the circle).

Lake effect rain event on the northern coast of Lake Erie which occurred in 1996. Photo from NOAA.

In meteorology and climate research, it is just as important to know when rain does not occur as when it does.  Therefore, when you check your rain gauge or snow board you are collecting data on how much rain or snow fell.  Noting zero rainfall in the rain gauge is as valid and important as noting that the cloud cover is “No Clouds” (0% cloud cover) or zero Hummingbirds were observed at the feeder.

Consider this: a desert is a place with little or no rain and is indicative of a location’s climate; a drought is a time when little or no rain falls, and is indicative of a longer term weather pattern.  Knowing when and where precipitation falls is significant in understanding the environment, as it helps to distinguish between whether or not a location is a desert or is only experiencing drought.  What this means in practice is that if you observe nothing in your rain gauge you should report zero for liquid precipitation.

If you see that there is no precipitation in the rain gauge, leaving that field blank is not the same as noting that no rain was in the rain gauge.  If you report zero, others will be certain that there was no rainfall and a zero will show on the map for your site.  Having your measurement of zero included in the day’s dataset improves the contours on the visualization and helps everyone recognize the true extent of GLOBE student observations and contributions to environmental knowledge.

Suggested activity:  Ask a friend, parent or teacher to describe the importance of zero in his or her daily activities.  Does zero have value in your life?  We hope through this post that you understand how valuable an observation of zero really is.  On your next visit to your atmosphere site, please be sure to take note of your rain gauge and report your rainfall amount, even if it’s zero.

Many records have been broken in the North Atlantic basin recently.  For example, the 2005 season saw a total of 28 named tropical systems, easily breaking the previous record of 21 storms.  Between the years of 2000 and 2012, 20 named tropical storms or hurricanes made landfall in the United States, and another 13 in Canada.  The map below shows the tracks of the landfalling storms from 2000-2011, as the 2012 storm tracks have yet to be entered into NOAA’s historical hurricane track database.

US and Canadian landfalling hurricanes from 2000-2011. Map courtesy of NOAA.

After the 2005 Atlantic hurricane season, many studies were formed to understand the National Oceanic and Atmospheric Administration’s (NOAA) ability to forecast these types of storms.  The Hurricane Intensity Research Working Group was established in order to document methods for improving hurricane intensity forecasting, an area of forecasts in which little improvement had been made in many years. Based on their recommendations, the Hurricane Forecast Improvement Project was established in 2009.

Some of the goals of the project are to reduce hurricane track and intensity errors, increase the probability of detection for rapid intensification of storms (rapid intensification is defined as the dramatic decrease of sea-level pressure over a short period of time), and extend the lead time of forecasts.  In order to reach these goals, scientists from the hurricane research, hurricane development and hurricane operations (also known as the forecasters) are working together to make improvements.

Projected path of Hurricane Sandy, the most recent U.S. landfalling hurricane. The cone indicates the probable, but not guaranteed, path that the storm could take. Photo courtesy of the National Hurricane Center

Scientists working on the project are hopeful that using new data assimilation systems and ensembles will greatly improve forecasts.  Data assimilation is the process of putting real observations into a weather prediction model, and ensemble forecasts is a technique where model forecasts are combined to generate a more accurate composite forecast.  Data from satellites and aircraft are also expected to have an impact on forecast improvement.

By improving hurricane forecasts, it is expected that there will be higher confidence in the forecasts so the public is more apt to respond saving lives and property.  It will be interesting over the next few years to monitor this project and see how hurricane forecasts are improved.

Suggested activity: Do you live in an area prone to these types of storms and experience the weather associated with them?  How could you use GLOBE protocols to understand more about them?  Let us know about it by leaving a comment here or on Facebook, or sending us an email at science@globe.gov.