Dams provide us with hydro-energy, and although in recant years there have been a few movements to remove dams, this is not the best solution in every case. Taking down dams requires a lot of energy and releases a lot of CO2 into the atmosphere as they are megalithic structures, not to talk about the consequences to the near by environment during and after this process. A much more cost beneficial solution in terms of both money and environment is to maintain the dams properly. However, the only dams which are monitored are the large dams. But sometimes this doesn't help, as monitoring and maintenance are not synchronized, like in case of recently busted Vale dam Brazil. And we have approx 68 dams fail each year in China, and the actual number of small and medium dams is unknown.
Sedimentation does not only effect the dam stability but also the hydro electricity production severely obscuring it, as does it effect the fresh water availability and future forecasting of agriculture and crops.
Therefore predicting any dam, energy, and water security is obscured by the lack of data, further more even the data that we have is incorrect as the sedimentation is a natural phenomena that occurs in every lake/reservoir. This is not being monitored for each dam, and there are no remote sensing applications addressing this issue as far as we are aware of it. Why it is important, global south, especially in N.Africa dams are constructed, operated and decommissioned via foreign aid programs, so their maintenance is not the best. Being able to remotely measure when it is the time to clean them, would be of crucial benefits for the local population depending on freshwater from that same source for irrigation and etc.
We want to develop a tool able to predict the levels of sedimentation, dam safety and freshwater availability worldwide and tell us which lake/reservoir has how many years of active usage before full sedimentation. To do this, we need to create a data set labeled by domain experts on the lake/reservoir sedimentation, first focusing solely on the deep vision neural network algorithm to be able to make a difference in colors as expert; once successful (second) introducing a third dimension, adding numerical data sets such as a contour line map of terrain, height of a dam,etc. training it to calculate a volume based on these, and comparing with ground data. Third predicting sedimentation volume based on different colors and volumes. The final step being introduction of fourth dimension which is time, and giving the algorithm the possibility to accurately predict the future freshwater availability and sedimentation rates within reservoirs, giving us opportunity to act in time.
There are to the best of our knowledge no entities working on this issue. No project we know of have tried to use remote sensing with deep vision to predict the freshwater and sedimentation. However there have been different tries to monitor sedimentation remotely, and predict freshwater availability like the HydroSheds project of WWF, EOMAP remote monitoring different water qualities, deepplanet looking into earth using deep vision, we are teh combination of all three. We would be essentially building up on these projects with the use of deep vision. This is why we call our project DamSmart, as it combines all the previous research on the subject of remote sensing of sedimentation, TSS and deep waters into one approach utilizing the advancements accomplished in deep vision and the availability of high resolution satellite imagery. Giving the world for the first time an accurate picture of dam safety, hydro power prediction, and freshwater availability worldwide.
We see ourselves in 5 years as an open source platform, which can tell every local municipality if their dam is endangered in the next 4 yrs and they need to budget for the dredging operations, if this sediments can be used in that region to stabilize land and local river basin and/or reservoir shores, if there will be hydro power shortages- allowing for better electricity management, tell every farmer if and when his crops will be endangered, allowing them to plan drought resistant crops, helping NGOs around the world that are working in drought ares to better cope and plan their operations in regards to water supply. Help UN secure their investments in dams and maintain them properly in case of war torn countries. Prevent future water wars,as all the information will be openly shared amongst internet users.
And the most important for our climate, all these sediments can be used to create nutrient dense topsoil in semi arid and arid regions, where dam reservoirs stability is not the best due to the lack of vegetation. Using sediments to create topsoil in days instead of10 yrs, as it is the case today, planting agroforestry can tremendously help rise the resilience of the local population and mitigate the climate change, by sinking carbon with trees, while preserving the dam stability and increasing the hydro power production.
To achieve all this we have partnered with multiple stakeholders. We are recognized by UNCCD * UN convention to combat desertification, as one of their NGOs. We have signed MoU with Dubai Municipality and are working on a pilot to demonstrate the use of sediments in re-greening desert, to stabilize the reservoir shores and green the surroundings, while increasing hydro power production. We are working with volunteers to create a connection between labelbox- a tool that we would prefer to use in labeling process and open source EU satellite data Copernicus Sentinel 2. We have a partnership agreement with BOKU University from Vienna, for 2 master programs questioning our methodology and trying to improve it. We are in process of making all our work available on github.
We hope to see your support in moving this very important and yet highly neglected issue forward!
If you have any questions