I can’t believe we’re in the holiday season. It feels like I took down the Christmas tree just a few weeks ago! In any case, it’s here, and with it will be some disruptions to my schedule. I have a lot going on next week so I won’t have a new blog post, but will resume the first week of December.
Last weekend I thought my post for the week would be a post-event analysis of a heavy rain event. Some of the models were indicating that a low pressure system aloft would be strong enough to trigger a heavy rain event over portions of the Big Island and Maui. There was considerable uncertainty in the event’s likelihood due to significant differences among the various models and their ensembles, but I thought there was a decent chance for some needed drought relief, especially for parts of the Big Island that had the highest rainfall potential. The loop below shows water vapor sector images from Tuesday morning on November 18. The trough aloft is west and southwest of the main Hawaiian Islands, and a deep layer of south to southwest winds are above 10,000 ft over the Big Island.
NOAA/NESDIS GOES water vapor image loop from 7:50 AM HST to 9:40 AM HST, November 18, 2025. The main Hawaiian Islands are left of center. This type of imagery shows water vapor in roughly the upper half of the atmosphere and is useful for tracking upper tropospheric features.
Ultimately, the low pressure system aloft was not strong enough to induce a surface trough as suggested by some of the model runs, and was not able to produce much in the way of rainfall. The 48-hour rainfall totals estimated by the NOAA National Severe Storms Lab’s Multi-Radar/Multi-Sensor (MRMS) system ended up being about 1 to 3 inches over the windward slopes of the Big Island and Maui, and mostly less than an inch over the rest of the state. The only leeward area to receive significant rainfall was the Kaʻū District on the Big Island. Several of the Kaʻū rain gage sites reported more than half an inch from the event. While helpful, they’ll need more to fully recover from the summer dryness. For most of the remaining leeward areas, the drought lives on.
Screen capture of the estimated rainfall for the 48-hours ending at midnight, November 20, 2025. The data display is from the NOAA National Severe Storms Lab’s Multi-Radar/Multi-Sensor (MRMS) system website. The image was cropped for visual clarity.
Instead of going into all kinds of glorious details about a heavy rain event, it looks like I’ll be talking about measuring rainfall for the rest of this post. Precipitation data are a fundamental component of both meteorological and hydrological observations. In Hawaiʻi, most of the land area receives precipitation solely as rainfall and not in any frozen form so I will only be discussing the measurement of rainfall here.
At a basic level, rainfall observations can be separated into two bins, manual and automated. In this post, I’ll just be covering manual observations, and will save the automated discussion for another time. A manual rainfall observation is just that, a manual reading by a person of water depth in a calibrated rainfall collector. The observation is recorded on a log sheet, on a computer, or both.
The National Weather Service (NWS) has a nationwide network of manually recorded rainfall observations under their Cooperative Observer Program (COOP). In the Hawaiian Islands, this network was quite extensive with over 300 active rain gages across the state in the mid-1990s. Sugar and pineapple plantations were excellent observers in this effort because they needed the data, could place gages at many locations in their vast properties, and had the staff to take the readings. The NWS provided and maintained the gages and associated equipment, and published the data in official publications. Having the official certification of the data by the federal government was beneficial for some cooperators, especially when applying to programs for things such as drought assistance.
Over time, the number of observers in Hawaiʻi has decreased substantially to less than a quarter of the number of active observers available in the mid-1990s. Some of the reasons for the drop in observers include the closure of the last sugar and pineapple plantations, and a general decrease in interest among private citizens interested in long term participation. A decrease in NWS organizational emphasis in COOP must also share some of the blame. In the past, the NWS had Cooperative Program Managers (CPM) who worked full-time on COOP duties. These individuals had the time to find observers and develop lasting relationships with them to help ensure site longevity and quality. Over time, organizational priorities changed and CPMs were phased out. The NWS now has Observation Program Leaders (OPLs) who have to handle COOP functions as an additional duty along with other tasks. COOP duties are sometimes assigned to junior staff members who move on when promoted. This prevents the development of lasting relationships with observers in the field and reduces community interest in the program.
Map of current NWS COOP sites in Hawaiʻi with manually observed rainfall, based on updated data in xmACIS. This is less than a quarter of the number of sites available in the mid-1990s.
Partly offsetting the reduction in NWS COOP sites is expansion of the Community Collaborative Rain, Hail and Snow (CoCoRaHS) network. CoCoRaHS is a non-profit organization with over 10,000 daily observations from volunteer participants across the United States, Canada, and the Bahamas. Observers enter their data online where the values are saved to a database and available shortly afterward on an interactive map. In Hawaiʻi, the number of participants has been slowly growing over the years, with the Big Island having the highest number of active participants. I started participating as a CoCoRaHS observer the day after I retired when I would be able to provide consistent observations in the morning. The map below shows Big Island CoCoRaHS sites and their rainfall totals for October. CoCoRaHS doesn’t completely offset the loss in NWS COOP sites, but it does provide quality data in some key locations.
Map of Big Island CoCoRaHS sites that provided rainfall data in October 2025. This screen shot is from the CoCoRaHS interactive map.
Participants in NWS COOP are provided a rain gage that is either an 8-inch diameter metal or a 4-inch diameter plastic tube. The image below shows an example of an 8-inch diameter NWS Standard Rain Gage (SRG) in the field being inspected by the Honolulu Forecast Office’s (HFO) CPM (now called the OPL). Regardless of type, each rain gage has an inner tube with a much smaller diameter to vertically extend the collected volume of water, which makes smaller amounts readable. This is necessary because a hundredth of an inch of rain is considered to be a measurable amount. Without the inner tube, it would be impossible to accurately discern a hundredth of an inch of collected rainfall. An example of an inner collection tube is included below. In this case, it’s a clear plastic tube that goes within a 4-inch diameter outer tube. Calibrated tick marks are printed on the inner tube for easy reading. In the case of the SRG, a specially calibrated measuring stick is used to determine the amount of rain collected.
Example of an 8-inch diameter NWS SRG in the field. The Cooperative Program Manager in the center and the observer on the right hold specially calibrated measuring sticks used to determine the amount of rainfall collected by the gage. It works just like an oil dipstick for your car.
Example of an inner collector tube from a plastic 4-inch diameter rain gage. Note the tick marks printed on the tube to facilitate easy determination of the rainfall amounts. This particular brand of gage also has amounts in millimeters printed in addition to inches.
The inner tube in the SRG type of gage can hold water equal to 2 inches of rainfall. Higher amounts of rain will overflow into the outer tube, which can hold 20 inches of rainfall. When the inner tube overflows, it is emptied out, and water collected in the outer tube is poured carefully into the inner tube (for 2-inch increments) as many times as necessary to completely measure all of the collected rainfall. Measurements for plastic 4-inch gages are similar, except that the capacities of the inner and outer tubes vary depending on the model of the gage.
Placement of the gage is extremely important for accurate rainfall observations. For COOP gages, we worked with the observer to find a location that was convenient for reading the gage, and free of nearby obstructions that would interfere with rainfall collection, such as a tree or building. In Hawaiʻi, the places with the most rainfall are often places with heavy vegetation coverage. While it may be difficult to avoid trees, you try as much as possible to find a spot that has the least amount of obstructions. You also have to consider future vegetation growth, and avoid placement near trees that may grow to interfere with rainfall collection years into the future.
You may be thinking that manually recorded rainfall is obsolete in this era of automation. While the availability of automated, real-time rainfall data is extremely useful, there is still a place for manually recorded rainfall readings, especially since there is no perfect automated system. One of the biggest advantages of a manual rainfall observation is that it is very difficult to beat the accuracy of a reading from a diligent observer. Automated systems all have issues that can be sources of errors or affect data availability. Manual observations can be used as a trusted source of data to validate automated systems, or fill gaps due to data outages.
Among the disadvantages of manual observations, one of the main issues is the limitation of one reading per day. This is particularly true in Hawaiʻi where observations of short duration and high intensity rainfall are crucial for monitoring flash flood events. One data point per day just doesn’t cut it. Another potential disadvantage is that by its nature, a manual observation will include the possibility of human error. This usually occurs in the entry of data onto log sheets or online. For example, an observation of 0.01 inches of rainfall can mistakenly be logged as 0.10 inches, or even 1.00 inches. Observations can also be logged on the wrong day. In one noteworthy example, an observer recorded rainfall on the wrong date for years before it was detected. Rather than logging rainfall at 7 AM on the date of the observation, he logged it on the previous day because that was when he felt most of the rain occurred. Stand-alone quality control checks would have missed this, but when the data were included during an analysis of heavy rain events, it was found that the large amounts were displaced by one day from all the other sites in the area!
If you’ve reached this far, congratulations! You can probably tell that I could go on quite a bit more when talking about rainfall observations. Initially, I was going to cover the basics of both manual and automated rainfall observations. I quickly realized that I needed to split them into separate posts because making one extremely long post would put you all to sleep, if you haven’t hit that point already. So I’ll stop now. I’m hoping to cover automated rain gages in December, assuming there aren’t too many heavy rain events to talk about in the coming weeks.
In closing I just want to say HAPPY THANKSGIVING!! I have many things to be thankful for this year, among them is your support of my blog!