Observers send thousands of precipitation reports. Then what happens to them? 

April 2024 - Nayoung Hur

At some point, the question may have crossed your mind of what happens to your precipitation observation after you press “Send It!”

To date, Mountain Rain or Snow observers have contributed over 25,000  observations in the 2023-2024 winter season, and over 65,000 observations since the project launched in 2020. Your simple submission allows our team and other scientists to better understand the rain-snow transition in both the short and long terms.  

all_obs_20240405 (1).png
obs_count_allyrs_20240405 (1).png

Heat map showing phase observations to date (left) and bar graph of total observations over the course of the project (right). TRoS = Tahoe Rain or Snow. The number of observations has increased each year since the beginning of the project, and observations come from many areas across the United States.

But before we can use the data from Mountain Rain or Snow, we follow a multi-step process that enables scientists to answer research questions with the data. This is akin to the steps needed to process freshly harvested wheat berries into flour so that it can be made into bread. Instead of a baguette or a bagel, however, the final output is what we call “the processed dataset”. This season, we made huge advancements in this process.

You submit your data, now what?

When you send an observation, it arrives in our database with your report of the precipitation phase (rain, snow, or mixed), as well as the date, time, and location from which the observation was made. Each observation is first quality controlled to make sure there are no mistaken reports. 

The next step is the biggest part of the process: We pair each observation with complementary information from other data sources to provide scientific context for each report. This step is necessary to answer Mountain Rain or Snow’s research questions in the same way that adding water and yeast to flour are needed to make bread dough. For example, we match each observation with the elevation from where it was submitted so that we can learn about rain-snow lines in each storm.

We elaborate on the types of information that we pair with each observation next.

The processing pipeline enables scientists to see the big picture 

The types of information that we pair with each observation are organized in three categories: geographical data, observed meteorological data, and data products.

A graphic showing a snowy mountain range with a marker indicating an observation from near the top of a mountain. Symbols of weather data adjacent to the marker indicate the complementary and ancillary data that are added to the raw data points during data processing.

Visualization of the information added to the raw data points during data processing.

The geographical data help us analyze your observations in terms of where they are located. For this, we need information on elevation, ecoregion, and state.

Meteorological data are also compiled for each observation. These include air temperature, dew point temperature, wet bulb temperature, and relative humidity. 

Finally, we compile estimates of precipitation phase from other sources in order to assess the accuracy of these technologies against your ground-truth observations. NASA satellite products, such as the Global Precipitation Measurement Mission, output a “Probability of Liquid Precipitation” estimate of whether precipitation will fall as rain or snow. 

*Want to get technical? See the “Epilogue” below for the sources of each of these types of information.

Why does this data processing pipeline matter?

Previous seasons’ efforts to process all of this information would take weeks of semi-manual processing and this would usually happen once a year, at the end of each season. This season, the Mountain Rain or Snow team optimized the workflow that we just described to process all of the observations daily. We wrote code in the form of an “R-package” (we named it rainOrSnowTools) to automate this process. The results are stored in a larger dataset from all previous seasons’ data equating to over 65,000 processed observations. 

process_report_back_whitebackground (1).png

Generalized schematic for the updated processing workflow.

Here's an example of what we can do once the data have been quality controlled and processed with all the complementary information. 

One of the big questions that our team endeavors to answer is: What is the probability of snowfall in each region? To answer this, we run analyses on the processed dataset and create snowfall probability curves. Across all years of this project, we have seen changes to the snowfall probability curves as we got more observations. For example, see the yearly changes for four ecoregions. The grey lines are each season’s data (delineated by water year), and the black line indicates the cumulative snowfall probability along with the 50% snowfall probability air temperature. Learning about these differences is critical for many reasons, one being that the air temperature at the 50% snowfall probability varies across regions. With more observations we can fine tune this value, which helps with winter weather forecasting, emergency management, and overall recreational safety. Our optimized data processing pipeline increases the frequency and reproducibility of analyses like this. 

snowfallprob_4locs_2024 (1).png

Snowfall probability curves over time, derived using precipitation phase report from MRoS observers. 

Data processing is just the beginning

Here are the main takeaways: Your contribution to Mountain Rain or Snow is the crucial ingredient in the scientific recipe for informing decisions and increasing knowledge about winter weather, and our newly optimized data processing approach allows for more efficiency so that these data can be used by researchers much sooner.

Teaser for future work: The processed dataset will be available through our dashboard that is currently in development**. With this feature, you will soon be able to visualize, subset, and download data from all years of the project.

*Epilogue: Want to get technical? 

Here are the sources of the complementary or ancillary data that we mentioned above: 

The elevation data and observed meteorological data are inputs to model meteorological variables for an observation point. If you are interested in more details about our modeling approach, we expanded on our methods and protocol in Arienzo et al. (2021). The processed dataset includes the raw observation data and all the possible modeled meteorological variables for the observation. 

**We are partners with several operational forecast offices who are using the data in real time to regionally view what is falling from the sky right now. If you are interested in using our data before the dashboard launch, please contact any of our team members to get connected. If you are an operational forecaster and would like to receive access to the real-time data, please reach out to Sonia Nieminen (  


Arienzo M.M., Collins, M., Jennings, K.S. (2021). Enhancing Engagement of Citizen Scientists to Monitor Precipitation Phase. Frontiers in Earth Science, 9. 

Mountain Rain or Snow logo with graphics of mountains, a snow crystal, and a rain drop.