Unprecedented crop loss due to 2019 Midwest US flood quantified from Space
Abstract
The Midwest US, which supplies a quarter of the world's grains, experienced a significant flood beginning in mid-March 2019. This event is the latest in a series of significant floods in the USA in the 21st century, turning fertile Midwest lands into marshes. To identify the hidden patterns in frequency and intensity of flood events, we investigate ~500 stream gauge stations with 50+ years of continuous data across the Midwest. We find that among the stations with statistically significant trends in each season, 61%, 69%, 80%, and 87% show a monotonic increase in the magnitude of the maximum seasonal discharge (or flooding) during spring, summer, autumn, and winter respectively. Such a dramatic increase in seasonal discharge over the past 50 years coincides with and is likely a consequence of, accelerated change in the climate.
Depending on the state, early to mid-June is the latest that corn and soybeans can be planted. On May 28, the USDA reported that only 58% of corn and 29% of soybeans that could be planted are in the ground, resulting in a catastrophic reduction in harvest. Many farmers voiced their concerns on social media using the hashtag #noplant19. To map the flood extent, we use ~550 synthetic aperture radar images acquired by Sentinel-1 satellite over the Midwest during Jan-May, 2019. We apply a supervised detection method that uses Bayesian statistics to calculate the conditional flood probability for each pixel. We find a significant number of farms along the Mississippi and Missouri rivers and their tributaries were flooded leading to the June deadline. These results are further validated using the Moderate Resolution Imaging Spectroradiometer mission monthly Normalized Difference Vegetation Index. Compared with the land use/cover data, we anticipate a substantial decline in 2019 grain harvest and export from the region. In the era of climate change, extreme precipitation is becoming more frequent, exacerbating flood events across the world. The data provided by Earth-observing radar satellites, such as Sentinel-1, provide an all-weather day-night mapping capacity that can be used for near-real-time assessment of flood hazards. Thus, they enhance crop monitoring and risk management efforts, which is crucial information for decision-makers and insurers to evaluate losses and compensate farmers' claims.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFM.H14G..01S
- Keywords:
-
- 1821 Floods;
- HYDROLOGY;
- 1833 Hydroclimatology;
- HYDROLOGY;
- 1840 Hydrometeorology;
- HYDROLOGY;
- 1855 Remote sensing;
- HYDROLOGY