Enhanced capability to monitor drought using citizen weather stations

Slides used during our presentation at the EGU 2021 conference, in the session "HS4.2 - Drought and water scarcity: monitoring, modelling and forecasting to improve hydro-meteorological risk management ". You can check the information of the session in this link: https://meetingorganizer.copernicus.org/EGU21/session/39888#

Transcript

1. Enhanced capability to monitor drought using citizen weather stations I. Garcia-Marti, M. de Haij, H. Leijnse, JW. Noteboom, A. Overeem, G vd Schrier vEGU21, Session HS4.2 Drought and water scarcity: monitoring, modelling and forecasting to improve hydro-meteorological risk management 30th April 2021
2. Background › Global warming: triggers or expands drought conditions at our planet › Netherlands: past several years we experienced drought conditions not measured since 1976. › KNMI: developed a cloud-based product showing the daily drought levels across country (2019): – Standard Precipitation Index (SPI) (McKee et al., 1993) – ~70 years of rainfall (1951-2010) › Present: Help decision-makers to assess meteorological drought and enable preventive measures
3. Background › Third-party data: weather observations measured by citizens! European met offices are interested in 3PD observations. › Citizen weather stations: citizens monitoring weather in private spaces (e.g. home, schools) using semi-professional devices › KNMI: Partner of WOW initiative (UK) since 2015, 900+ CWS in the network, 150+ M observations (2015 -2020) › Present: we studied the quality of temperature, wind speed and rainfall observations WOW-NL KNMI
4. Research questions What if we combine the SPI with WOW observations? › What is the added value of 3PD observations (e.g. research, operational, economic) provide wrt official networks? › Assessing the possibility of providing higher resolution drought maps including WOW-NL data › Enable the possibility for underrepresented regions to obtain local drought metrics WOW-NL rainfall (2018-2019) daily gridded precipitation (1951-2011) SPI Calculations Calibrated model 5-min interval aggregation Quality control QC’d WOW- NL rainfall Radar data at 5-min interval SPI-WOW KNMI’s SPI DROUGHT INDEX SPI-WOW DROUGHT INDEX
5. 5 After QC, results show moderate positive correlation ☺ But who are these dry stations?
6. 6 Random distribution suggests “faulty stations”… that do pass the QC
7. 7 Overlay official SPI vs WOW SPI Difference between two networks Visualizing skill (GIF animation, should move ☺) (colors should blend) (difference should be grey)
8. 8 Visualizing skill ~30km Good agreement between official and WOW SPI Is it a glitch of the WOW network, or is it that the official SPI is underestimating local drought conditions?
9. Recap WOW-NL picks drought signal Pinpoint local drought conditions A thorough comparison is needed Value 3PD might contribute to official products! Research: Hi-resolution indicators of drought risk Operations: underrepresented areas -> local drought Big picture Multi-Hazard Early Warning Systems Hi-res weather warnings and forecast Climate Services Wrapping up
10. Questions? ☺ garciamarti@knmi.nl Thanks! 10