Machine Learning for Location Low Head Dam

Low-head dams (LHD) also known as “drowning machines,” “keepers,” and “killer weirs” are structures build across stream channels that raise the water level upstream to allow for diversion for irrigation or consumption purposes. The phenomenon that can makes the downstream water (tailwater) dangerous is called submerged hydraulic jump. A submerged hydraulic jump forms when high tailwater levels drowns out the hydraulic jump that forms at the base of every low-head dam and produces a recirculating current. The force from this current is sufficient to trap recreationists with almost no chance of escape. Besides this threat to public safety, LHD also create barriers to fish and aquatic organism passage. Creating an inventory of low-head dams has become a priority as the number of drowning deaths increases at these structures(Kern, Hotchkiss et al. 2015)
One of the reasons why it is extremely difficult to locate existing low head dams is because they are not regulated as part of state dam safety programs. There is a national task force constantly making progress on creating a national inventory of low head dams by manually identifying them using Google Earth Pro. This process is time consuming, requires a lot of resources, and requires the help of experienced professionals on the topic to perform Quality Control on the collected data. (Continued)