High-Level Project Summary
Floods are the environmental phenomenon that affects more people around the world. Flooded Space is an application that aims to contribute to the solution of this problem, predicting the risk of floods through the use of meteorological data from satellite platforms and meteorological centers, focused on tropical humid regions of the Amazon River as a case study; and, additionally, implementing effective communication channels so that this information reaches government agencies, the affected population and relief agencies in a relevant way, to facilitate decision making.
Link to Project "Demo"
Link to Final Project
Detailed Project Description
"The population is increasing at a higher rate in areas we know have been flooded in the recent past. This exposure element is being driven, in part, by how we decide where to develop" - Jonathan Sullivan, postdoctoral scientist at the University of Arizona [2]
It is necessary to understand why people are moving to floodplains in order to be able to contribute to flood mitigation. Much of the population that has migrated to these areas is in Asia and sub-Saharan Africa, where at least 213 million people are at risk of flooding [1]. The characteristics that most define these populations are the difficulties due to low income conditions, low accessibility to technologies, low level of use of information and therefore less capacity to deal with potential disasters [2].
Floods are the environmental phenomenon that affects most people around the world. During the period 2000-2015 more than 255 million people were affected at least once by major floods. The picture indicates that the proportion of the population exposed to floods will increase further in the coming decades [1].
its implementation Flooded Space is divided into five stages:
1:
Systematize each country’s spatial data observatories, land uses, since they are a basic tool for planning. While in many countries use maps are outdated, it is essential that they can be updated, as they are the responsibility of the regions of each municipality or province.
At this stage the whole analysis will be focused on the basin, since it represents the intrinsic characteristics of each region (precipitation), outside the administrative limits of each region, municipality and/or country. The focus is on land use and the basin, since in the following stages it will be possible to extract information from the flood and overlay the land use units, to then generate statistics and areas affected by each type of land use, with the objective that the actors can make decisions based on these analyses.
Stage 2:
For data acquisition we use the ESA database on the Sentinel -1 SAR GRD, which operates in the C band, offers good spatial resolution and is available on the ESA site [3], then we filter the search with the polarization that gives us better visualization of the areas where there might be water.
Below we take a range of dates to filter the data on our case study which will be a flood in the department of Arauca Colombia occurred on July 13, 2021, We will use two ranges, the first describes the area without flood and the second is a range that describes the time after the flood occurred, we also choose the direction of the orbit in which the image was obtained.
Because the radar image presents a speckle phenomenon, a suitable filter is applied and the image is smoothed for better viewing.
Below we make use of the data obtained in step 1 on the digitization of the land use units that helps us to choose areas with visibly differentiable characteristics in the radar image, which correspond to forest, permanent water urban area among others, finally we made a comparison of radar images to determine which areas are flooded, It should be noted that this subsequent process is automated and is carried out in an unsupervised way so the end result is the same application but to determine floods in any area at global level. It also includes mapping
The next step is to use radar polarizations which make it easier to visualize flooded areas, especially the polarization with which we got the best results was (VV).
We then determined the training data which were the selected characteristics and the polarization to use, then used the samples to train an artificial intelligence sorter called Gradient Boost, which generates a predictive model from the model set found. We verify the errors obtained by the classifier and evaluate the result.
Finally we get a layer with the display data corresponding to the flooded areas indicated in blue.
Develop Code: https://earthengine.googlesource.com/users/adrrod44/Hakaton_Nasa
Stage 3:
The relationship between rainfall and floods is a concept that is popularly internalized, so it is clear that the use of precipitation data is useful and even vital for the implementation of the flood prediction and prevention system. The source of this data can be from both satellite sensors and the stations of the meteorological centers, since both complement and corroborate each other, so the use of Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) is useful an open set of quasi-global precipitation data that incorporates our in-house climatology, CHPclim, 0.05° resolution satellite imagery, and in-situ station data to create gridded rainfall time series. [4]
Stage 4:
Nowadays the information of the spatial type, tabular, graphic, etc., is important for the decision making whatever its purpose and even more when we talk about monitoring flood areas, where the level of information needs to be transparent and timely for all beneficiaries and for the definition of public policies. This level of information has to reach all levels, from the beneficiaries (Communities), Municipal technicians to the Central Level (Ministry) with the purpose of empowerment of information and problem.
In this sense, the use of the Internet as a channel for the transmission of geographical information facilitates the exchange of spatial data between the different agencies, in addition to promoting "interoperability", avoiding duplication of information and consequently, improving intersectoral coordination of geographic information production programmes.
The approach proposed by the Consulting Team is a REACTIVE PLATFORM, meaning that parameters will be defined that can change depending on the elements of the viewfinder and this will allow the Map or the graphic to change depending on the needs of the User. This approach is now innovative in that it makes it possible to understand the dynamics of the problem and to improve decision-making.
Stage 5:
As can be seen from the outline, the objective and fundamental part is that the proposed tool is actually useful to the actors. In this sense two types of actors are identified. The first are technicians of the Municipalities and therefore those in charge of having the relevant information through the viewfinder and being able to send the message via Radio system. We have seen in Latin America that although the use of ICTs in this case the cell phone, is growing as well as coverage, however radio is still a device that is most used especially in rural areas (75% in Bolivia).
Communities are another actor, although they have mobile devices, but they do not always have an internet connection, that is why it has been seen that the team can use a platform that connects with the System and can generate alert thresholds such as precipitation, and to be able to send this information in a preventive way to the communities. The information will be delivered to the population through alerts via SMS messages, and each alert will contain the level of risk within the hydrographic basin and also an access link to the platform to view the information on the map. This aspect, beyond sending information, will also have a component where communities can be part of the solution, in the sense that they can also send messages or photos of possible flood risks. This process will have two validation inputs, on the one hand the analysis carried out in Stage 2, and this analysis of validation by the actors, as well as achieving the empowerment and strengthening of educational units with the use of the platform (Training).
Space Agency Data
Hackathon Journey
An amazing experience sharing with partners from Ecuador that joined us to tackle this ongoing problem in Colombia. Physics, electronics and geographical engineers in the same spot sharing knowledge and matching togheter their actual research and state projects to find better solutions!
References
[1] Tellman, B., Sullivan, J.A., Kuhn, C. et al. Satellite imaging reveals increased proportion of population exposed to floods. Nature 596, 80–86 (2021). https://doi.org/10.1038/s41586-021-03695-w
[2] https://earthobservatory.nasa.gov/images/148866/research-shows-more-people-living-in-floodplains
[3] https://sentinel.esa.int/web/sentinel/sentinel-data-access
[4] https://www.chc.ucsb.edu/data/chirps
[5] Carreño Conde, F., & De Mata Muñoz, M. (2019). Flood monitoring based on the study of Sentinel-1 SAR images: The Ebro River case study. Water, 11(12), 2454.
Global Judging
This project has been submitted for consideration during the Judging process.