The AAU VIRADA dataset contains a total of 215 hours of surveillance video from two different locations in Aalborg, Denmark. The purpose of the dataset is to enable the use of surveillance cameras as surrogate rain gauges.

The ground truth precipitation data comes from two disparate sources: 1) mechanical, tipping-bucket rain gauges and 2), an advanced laser disdrometer.

The footage from the first location, Crossing1, is split into a training (trn) and a validation (val) split. The footage from the second location, Crossing2, is used entirely for testing.

More details on the setup is found in the CVPR Workshop paper:

Haurum, Joakim Bruslund, Chris Holmberg Bahnsen, and Thomas B. Moeslund. "Is it Raining Outside? Detection of Rainfall using General-Purpose Surveillance Cameras." 2019 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). IEEE, 2019.

This dataset contains 32072 daily time series showing the temperature observations and rain forecasts, gathered by the Australian Bureau of Meteorology for 422 weather stations across Australia, between 02/05/2015 and 26/04/2017.

The original dataset contains missing values and they have been simply replaced by zeros.

The dataset contains records from the Town of Cary, NC Internet-of-Things rain gauges. This dataset contains 60 days' worth of data.Used as a part of the Stormwater Monitoring Dashboard

These data include daily precipitation measurements from nine different recording rain gages (RRG) at Coweeta Hydrologic Laboratory in Macon County, North Carolina, USA. These stations are operated by the Southern Research Station, USDA Forest Service. Data include total daily precipitation for the following recording rain gages: RRG05 (1992-2017), RRG06 (1936-2017), RRG12 (1942-2017), RRG13 (1942-2017), RRG20 (1962-2017), RRG31 (1958-2017), RRG41 (1958-2017), RRG55 (1990-2017), and RRG96 (1943-2017).

Rain

## Overview

Rain is a dataset for object detection tasks - it contains R annotations for 476 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Real Rain Dataset

## Overview

Real Rain Dataset is a dataset for object detection tasks - it contains Objects annotations for 992 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

and multi provision of appropriate predictions are created in real-time similar resolution. Then

CHIRPS is an abbreviation for Climate Hazards Group InfraRed Precipitation with Station Data (Version 2.0 final). The CHIRPS is a 30+ year quasi-global rainfall dataset and incorporates 0.05° resolution satellite imagery with in-situ station data to create gridded rainfall time series for trend analysis and seasonal drought monitoring. The data of the CHIRPS pentad is derived from Google Earth Engine with earth engine snippet as https://code.earthengine.google.com/?scriptPath=Examples%3ADatasets%2FUCSB-CHG_CHIRPS_PENTAD . With the dataset in a global format, it is clipped with the Cambodia boundary and generated the data visualized chart through the obtained data.

In 2024, the United States saw some **** inches of precipitation. The main forms of precipitation include hail, drizzle, rain, sleet, and snow. Since the turn of the century, 2012 was the driest year on record with an annual precipitation of **** inches. Regional disparities in rainfall Louisiana emerged as the wettest state in the U.S. in 2024, recording a staggering ***** inches (*** meters) of precipitation—nearly **** inches (ca. ** centimeters) above its historical average. In stark contrast, Nevada received only **** inches (ca. ** centimeters), underscoring the vast differences in rainfall across the nation. These extremes illustrate the uneven distribution of precipitation, with the southwestern states experiencing increasingly dry conditions that experts predict will worsen in the coming years. Drought concerns persist Drought remains a significant concern in many parts of the country. The Palmer Drought Severity Index (PDSI) for the contiguous United States stood at ***** in December 2024, indicating moderate to severe drought conditions. This reading follows three years of generally negative PDSI values, with the most extreme drought recorded in December 2023 at *****.

Low Rain

## Overview

Low Rain is a dataset for object detection tasks - it contains Helmat Helmet annotations for 219 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

[Metadata] Mean Annual Rainfall Isohyets in Millimeters for the Islands of Hawai‘i, Kaho‘olawe, Kaua‘i, Lāna‘i, Maui, Moloka‘i and O‘ahu. Source: 2011 Rainfall Atlas of Hawaii, https://www.hawaii.edu/climate-data-portal/rainfall-atlas. Note that Moloka‘I data/maps were updated in 2014. Please see Rainfall Atlas final report appendix for full method details: https://www.hawaii.edu/climate-data-portal/rainfall-atlas. Statewide GIS program staff downloaded data from UH Geography Department, Rainfall Atlas of Hawaii, February, 2019. Annual and monthly isohyets of mean rainfall were available for download. The statewide GIS program makes available only the annual layer. Both the monthly layers and the original annual layer are available from the Rainfall Atlas of Hawaii website, referenced above. Note: Contour attribute value represents the amount of annual rainfall, in millimeters, for that line/isohyet. For additional information, please see metadata at https://files.hawaii.gov/dbedt/op/gis/data/isohyets.pdf or contact Hawaii Statewide GIS Program, Office of Planning and Sustainable Development, State of Hawaii; PO Box 2359, Honolulu, Hi. 96804; (808) 587-2846; email: gis@hawaii.gov; Website: https://planning.hawaii.gov/gis.

Person In Rain

## Overview

Person In Rain is a dataset for object detection tasks - it contains Person annotations for 1,000 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

1) Data Introduction • The Rain in Australia Dataset is a tabular weather forecasting dataset, including daily weather information collected for approximately 10 years from various weather stations across Australia, and next-day precipitation (more than 1 mm, RainTomorrow).

2) Data Utilization (1) Rain in Australia Dataset has characteristics that: • Each row contains a variety of daily weather variables and target variables (RainTomorrow: Next Day RainTomorrow) such as date, region, highest/lowest temperature, precipitation, humidity, wind speed, and air pressure. • The data reflect multiple regions and various weather conditions, making them suitable for time series and spatial weather pattern analysis and the development of binary classification prediction models. (2) Rain in Australia Dataset can be used to: • Development of precipitation prediction models: Machine learning-based next-day precipitation prediction (whether an umbrella is required) models can be built using various weather variables and RainTomorrow labels. • Weather Patterns and Regional Analysis: By analyzing regional and seasonal weather variables and precipitation patterns, it can be used to establish customized weather strategies for each industry, such as climate change research and agriculture and tourism.

rainrd/rain dataset hosted on Hugging Face and contributed by the HF Datasets community

An extensive precipitation database at the 149 km2 Walnut Gulch Experimental Watershed (WGEW) has been developed over the past 53 years with the first records starting in August 1953 and continuing to the present. The WGEW is a tributary of the San Pedro River, surrounds the town of Tombstone in southeastern Arizona, and has a drainage area of approximately 149 km2. Elevation of the watershed ranges from 1220 m to 1950 m above mean sea level (MSL). Average annual precipitation for the period of 1956-2005, as measured with six gauges, is roughly 312 mm, with approximately 60% falling during the summer monsoon. Precipitation consists almost solely of rainfall with relatively rare instances of hail and snowfall. From a historical high of 95 rain gauges, a current network of 88 gauges is operational. This constitutes one of the densest rain gauge networks in the world (0.6 gauges/km2) for watersheds greater than 10 km2. Through 1999, the network consisted of analog recording weighing rain gauges. In 2000, a newly designed digital gauge with telemetry was placed adjacent (1 m) to the analog gauges. Both the analog and digital networks of gauges were in operation from 2000 to 2005 to enable a comparative analysis of the two systems. The analog data were digitized from paper charts and were stored in breakpoint format. The digital data consist of rainfall depths at 1-min intervals during periods of rainfall. Resources in this dataset:Resource Title: GeoData catalog record. File Name: Web Page, url: https://geodata.nal.usda.gov/geonetwork/srv/eng/catalog.search#/metadata/WalnutPrecipitation_jjm_2015-03-20_1018

Weighing rain gauge charts record the amount of precipitation that falls at a given location. The vast majority of the Weighing Rain Gauge Recording Charts collection comes from the National Weather Service, cooperative stations, and the Pacific Islands, however, there are some forms that were provided through international organizations and agreements. Throughout much of the history of the collection, observations were taken on paper. However, much of the paper was converted to other media. Of those records that were converted, records are available on microfilm from 1888 through 1972 and on microfiche between 1973 and the mid-1990s. Data contains continuous recording of precipitation amounts to one-hundredth of an inch.

Hourly Precipitation Data (HPD) is digital data set DSI-3240, archived at the National Climatic Data Center (NCDC). The primary source of data for this file is approximately 5,500 US National Weather Service (NWS), Federal Aviation Administration (FAA), and cooperative observer stations in the United States of America, Puerto Rico, the US Virgin Islands, and various Pacific Islands. The earliest data dates vary considerably by state and region: Maine, Pennsylvania, and Texas have data since 1900. The western Pacific region that includes Guam, American Samoa, Marshall Islands, Micronesia, and Palau have data since 1978. Other states and regions have earliest dates between those extremes. The latest data in all states and regions is from the present day. The major parameter in DSI-3240 is precipitation amounts, which are measurements of hourly or daily precipitation accumulation. Accumulation was for longer periods of time if for any reason the rain gauge was out of service or no observer was present. DSI 3240_01 contains data grouped by state; DSI 3240_02 contains data grouped by year.

Water is an essential ingredient to life on Earth. In its three phases (solid, liquid, and gas), water continuously cycles within the Earth and atmosphere to create significant parts of our planet’s climate system, such as clouds, rivers, vegetation, oceans, and glaciers. Precipitation is a part of the water cycle, where water particles fall from clouds in the form of rain, sleet, snow, ice crystals, or hail. So how does precipitation form? As water on Earth’s surface evaporates it changes from liquid to gas and rises into the atmosphere. Because air cools as altitude increases, the vapor rises to a point in the atmosphere where it cools enough to condense into liquid water or freeze into ice, which forms a cloud. Water vapor continues to condense and stick to other water droplets in the cloud until the weight of the accumulated water becomes too heavy for the cloud to hold. If the air in the cloud is above freezing (0 degrees Celsius or 32 degrees Fahrenheit), the water falls to the Earth as rain. If the air in the cloud is below freezing, ice crystals form and it snows if the air between the cloud and the ground stays below 0 degrees Celsius (32 degrees Fahrenheit). If a snowflake falls through a warmer part of a cloud, it can get coated in water, then refrozen multiple times as it circulates around the cloud. This forms heavy pellets of ice, called hail, that can fall from the sky at speeds estimated between 14 and 116 kmph (9 and 72 mph) depending on its size. A hailstone can range from the size of a pea (approximately 0.6 cm or 0.25 inches) to a golf ball (approximately 4.5 cm or 1.75 inches), and sometimes even reach the size of a softball (approximately 10 cm or 4 inches).Precipitation doesn’t fall in the same amounts throughout the world. The presence of mountains, global winds, and the unequal distribution of land and sea cause some parts of the world to receive greater amounts of precipitation compared with others. Areas with rising moist air generally indicate regions with high precipitation. According to the Köppen Climate Classification System, tropical wet and tropical monsoon climates receive annual precipitation of 150 cm (59 inches) or greater. Tropical wet regions, where rain occurs year-round, are found near the equator in central Africa, the Amazon rainforest, and southern India. Monsoons are storms with large patterns of wind and heavy rain that can span over a continent. Tropical monsoon climates are located mainly in Southeast Asia and areas around the Pacific Ocean, where annual rainfall is equal to or greater than areas with a tropical wet climate. Here, intense monsoon rains fall during the three hottest months of the year, which are usually between June and October. Snow and ice, which are most common in high altitudes and latitudes, cover most of the Earth’s polar regions. High altitude regions of the Andes, Tibetan Plateau, and the Rocky Mountains maintain some amount of snow cover year-round.Over the next century, it is predicted warming global temperatures will increase the temperature of the ocean and increase the speed of the water cycle. With a quicker rate of evaporation, there will be more water in the atmosphere, allowing clouds to produce heavier precipitation and more intense storms. Although storms would be more intense in wetter regions, increased evaporation could also lead to extreme drought in drier areas of the world. This would greatly affect farmers who grow crops in dry locations like Southern California or Kansas.This map layer shows Earth's mean precipitation (measured in centimeters per month) averaged from 1981 to 2012 as calculated but the Copernicus Climate Change Service. The data was collected from the Copernicus satellite and validated with precipitation measurements from weather stations. Scientists averaged all of the amounts (originally collected in meters) occurring each month together, and they calculated the average of each month over 30 years to create this map.

What is NeRAIN?The Nebraska Rainfall Assessment and Information Network (NeRAIN) is a program designed to get citizens involved with monitoring weather across the state. The network consists of several hundred volunteers that spend a few minutes each day recording precipitation measurements along with outstanding weather events and uploading them to the NeRAIN website. Due to weather changing in time and space, many volunteers are needed in order to accurately measure and describe rainfall, snowfall patterns, and hail swaths across the state.Who is in charge of NeRAIN?The Nebraska Department of Natural Resources (NeDNR) operates the website. Rain gauges have been provided to NeRAIN volunteers using funding by the Nebraska Natural Resource Districts (NRDs), NeDNR, and the Nebraska Environmental Trust (NET). Volunteer training and assistance is provided by regional coordinators in each of the Nebraska NRDs.To find your regional coordinator, see the coordinator list at: https://nednr.nebraska.gov/NeRain/Home/CoordinatorsTo find your local NRD, utilize the following map: https://www.nrdnet.org/nrds/find-your-nrdWho can volunteer to join NeRAIN?NeRAIN is a community project. Everyone can help, both young and old. The only requirements are an enthusiasm for watching and reporting weather conditions and a desire to learn about the power and beauty of our natural world, along with the ability to upload data using a computer or smartphone. NeRAIN welcomes participation by any motivated individual, family, or groups such as schools or clubs.How does the program work?Volunteers are asked to read their rain gauge at 7 AM each morning. While 7 AM is the target, any time within a couple of hours before or after (5 AM – 9 AM) would still provide useful information. The rainfall amount received during the prior 24-hour period is then uploaded to the website via the volunteer’s smartphone or computer.Additional information may also be uploaded. During the winter volunteers may upload snow depth, or the amount of water in the snowpack on their property. Users may report hail or other intense storm activity. Other useful information could include the intensity of drought (presence of cracks in the ground, dry vegetation, etc.).Recording when it doesn’t rain it is just as important to report as when it does rain. Statistically, recording a zero for precipitation is equally important as recording a storm.What happens to the data?NeRAIN information is updated daily and is available for public access on this website. The data can be seen on an interactive map as well as in several reports. The data may be downloaded in table format.The data is also transmitted several times each day to the Community Collaborative Rain, Hail, and Snow Network (CoCoRaHS). CoCoRaHS is an international database for precipitation and weather events based in Fort Collins, Colorado on the campus of Colorado State University. The data provides important daily decision-making information for agriculture, industry, home water use, utility providers, insurance companies, resource managers, and educators. The data is used by the National Weather Service to ground truth weather radar returns and to fill in information between official weather stations.You may find out more about the program here: https://nednr.nebraska.gov/NeRain/Account/Involved

The European climatological gauge-adjusted radar precipitation dataset, EURADCLIM, addresses the need for an accurate (sub-)daily precipitation product covering 8000000 square kilometers of Europe at high spatial resolution. It consists of 1-h and 24-h precipitation accumulations (every clock-hour) at a 2-km grid for the period 2013 through 2022. It is based on the European Meteorological Network (EUMETNET) Operational Program on the Exchange of weather RAdar Information (OPERA) gridded radar dataset of 15-min instantaneous surface rain rates. For EURADCLIM, first methods have been applied to further remove non-meteorological echoes from these images by applying two statistical methods and a satellite-based cloud type mask. Second, the radar composites are merged with the rain gauge data from the European Climate Assessment & Dataset (ECA&D) in order to substantially improve its quality. We expect to rerun EURADCLIM once a year over the entire period, using all available ECA&D rain gauge data, and extend it with one year of data. This will result in a new version of this dataset. Project EURADCLIM was financed by KNMI’s multi-annual strategic research programme (project number 2017.02). The EURADCLIM dataset is based on the OPERA surface radar rain rate and daily precipitation sums of the rain gauge networks provided by the European national weather services and other data holding institutes, through ECA&D. With respect to version 1, the changes include slightly improved removal of non-meteorological echoes, somewhat better rain gauge coverage over the years 2013 to 2020, and years 2021 and 2022 have been added to the dataset. Usage: For each month a zip file is provided. The data are in UTC, where the time in the unzipped filenames is the end time of observation in UTC. Object "dataset1/data1" contains the 24-h precipitation accumulation in millimeters. For a given grid cell, it is equal to the "nodata" value (-9999000.0) in case data availability of the underlying 1-h accumulations is < 83.3%. Object "dataset2/dataset1" contains the number of valid values for each radar grid cell (count), i.e., the number of underlying 1-h accumulations that have been used.