View this email in your browser

  November Newsletter  

Cloud to Street is a global platform for mapping and monitoring floods and flood risk at high resolution and in partnership with local users. Eleven national governments rely on our tools to respond to disasters faster and reduce risks long-term. 

You're receiving this email because you signed up to learn more about us.
Get involved
Mapping Unmapped Floods During Crisis in the Congo

The Republic of Congo has been in a state of sustained crisis since extreme flooding brought on by heavy rains began in early October. On November 7th, the government issued an appeal for international assistance and humanitarian aid, citing the urgent needs of the affected population along the Ubangi River, even before rainfall on November 21 caused the flooding to become even more severe. The population in the Congo is particularly vulnerable to flooding; the World Food Programme Congo country office writes that "Congo has faced worse and worse floods for the past 20 years: an effect of climate change. Populations can no longer predict the weather, which affects the agricultural calendar, threatening food security."

Cloud to Street has a long-lasting partnership with the World Food Programme country office. In 2018,
as Planet reported in this piece, we provided maps that proved that refugee camps along the Congo River were at high-risk. We were ready to help however we could when WFP Country Director Jean-Martin Bauer, facing unsubstantiated reports that remote parts of the country were badly flooding, asked for assistance. 

"The WFP didn't have any clear reports from the government—there's just not a lot of capacity for flood monitoring on the ground,"
Remote Sensing Scientist Tyler Anderson explained. "They had a rough estimate of the number of people affected in one department, Likouala, and they didn't have an idea about two other departments, Cuvette and Plateaux."

Cloud to Street scientists were able to analyze that there were at least 20,000 victims of flood in Cuvette and Plateaux. In fact, in these three districts altogether we calculated that there were at least 70,000 vulnerable people—a staggering number, for whom the WFP can now coordinate much-needed humanitarian aid.

See our maps of the flooding in Likouala, Cuvette, and Plateaux regions above. Photo from WFP Congo.

Training in Nairobi Sparks Discussion of Climate Resilient Futures
After days of coding and mapping workshops with our partners at the Nile Basin Initiative and the World Bank, participating hydrologists and government representatives told us how the better data we can provide can unlock new possibilities in their work.

With our support, local stakeholders from Kenya, Tanzania, Burundi, and Rwanda can now rethink their flood planning and response. From establishing dynamic buffer zones for infrastructure, calibrating existing flood forecast models, understanding the paths of expanding wetlands and meandering rivers, mapping disasters otherwise too remote to access, to overcoming limited and aging hydromet stations — these projects would increase climate resilience and protect countless lives.
How Machine Learning Unlocks More Accurate Flood Predictions

Mapping, monitoring, and analyzing floods using remote sensing science, as we do here at Cloud to Street, has its challenges. Satellites—the very technology that makes it possible to remotely create scalable, low-cost, and high-quality flood maps and impact assessments—have structured revisit times that do not capture every part of the earth every day, and clouds and their shadows can block them from imaging the ground. 

Cloud to Street Machine Learning Scientist Derrick Bonafilia believes that artificial intelligence, namely deep learning, convolutional neural nets, and random forest models, holds the particular power to circumvent these satellite-specific challenges. Unlike traditional hydrological flood modeling, which is dependent upon physical stream gauges, elevation, and rainfall measurements, with satellites “there’s a lot of room for rethinking things in a more data-driven way. Current research barely scratches the surface,” Derrick explained. 

“With machine learning, the scope of [what you can do with satellite data] is much bigger,” Derrick said. “Machine learning algorithms can interpret different bands in the visual spectrum in a way that human eyes can’t, so you can see more things through clouds. For example, when dealing with LANDSAT [satellite] data, these algorithms can examine at 11 bands at once. It can see more relationships than we can, and with potentially with more accuracy.” For context, human eyes can see only 3 bands. 

After a month of leading supervised learning efforts, Derrick is already seeing results. Leapfrogging the inherent limitations of satellite data, the improved algorithm can, astoundingly, identify and visualize flooding under cloud shadows and in more varied atmospheric conditions.

“Machine learning is rapidly accelerating our product development. Within a couple of weeks, we were able to take the algorithm we developed for 3 years and increase its accuracy significantly,” Chief Science Officer Dr. Beth Tellman said. “In addition to beating our old algorithm under clear conditions, we are also seeing signs that we will be able to improve coverage in other conditions too. It’s very exciting.” 

How Big Data Can Democratize a Climate Safety Net
Bessie presents at TTI/Vanguard's [NEXT] conference. @Tommy-Reed draws her talk.
Find Us at AGU

Cloud to Street scientists are excited for their addresses at the American Geophysical Union's 2019 meeting in San Francisco. To connect with Chief Science Officer Beth Tellman and Director of Technology Colin Doyle, attend their talks and/or reach out via email.

December 9
  • 2:40 pm: Co-founder and Chief Science Officer Dr. Beth Tellman presents "From Publishable to Operational: New Metrics to More Honestly Measure the Ability of Remote Sensing Algorithms to Consistently Monitor Flooded Assets and Populations in Near Real-Time" (Moscone West -2000, L2).
December 11
  • 8:00 am - 12:20 pm: Dr. Tellman presents "Monitoring Inundated Infrastructure and Assets with High Resolution Satellite Imagery to Enable Financial Protection" (Moscone South - Poster Hall).
  • 8:45 am: Director of Technology Colin Doyle presents "Leveraging Earth Observations for Decision Support During Flood Disaster Support During Flood Disaster Prevention, Response, and Recovery" (Moscone South -301-302, L3).
Welcome our new team members
We are thrilled to be working with Dr. Marco Tedesco to push the boundaries of passive microwave remote sensing science to better map, monitor, and analyze catastrophic floods.
Read about his recent work studying accelerating ice sheet melting here. 
We want to hear from you.
Get involved
Let's stay in touch.
Copyright © 2019 Cloud to Street, All rights reserved.

Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.

This email was sent to <<Email Address>>
why did I get this?    unsubscribe from this list    update subscription preferences
Cloud to Street · 4938 McConnell Ave · Los Angeles, CA 90066-6714 · USA

Email Marketing Powered by Mailchimp