Researchers in India have recently implemented Artificial Neural Network (ANN) to predict groundwater levels in five aquifers in Prakash India. This system mimics human and animal neural response systems to analyze nonlinear and complex data that humans cannot with high accuracy. It uses a series of algorithms to analyze data about a certain topographical area and its watershed to forecast the outcome of different possible water usage situations.
The study was carried out in Andhra Pradesh in India, a region characterized by poor soil conditions, sporadic rainfall and little vegetation. Its main source of drainage is the massive granite hills to the north that disperse into many smaller tributaries. By inputting all this data as well as other factors such as differences in the slope and rainfall in particular areas they were able to create five year forecasts for different possible water usage outcomes. There are many different stages to ANN analysis. Firstly, all the data must be inputted into the system, wrong values and vastly different measurements are removed for accuracy. After this is completed, a series of algorithms are used to properly analyze up to 23 months worth of data. Many different solutions are created and the best possible route is chosen to optimize water use. The validity of the forecast is then validated and is carried out for different wells
This mode of research is an excellent way to learn about water stewardship. It is completely non invasive and allows people to understand just how much water there is available for use. The study estimated that if overpumping in one area can affect four different wells and that levels could hit zero even while the aquifers are in a state of recharge. They also predict that decreasing pumping rate could increase groundwater levels by eight to ten meters suring months of recharge.
So far in Andhra Pradesh, forecasting has been carried out for five geologically different wells for up to five years in advance. The changes from year to year are predicted by hypothetically creating different water usage strategies and this mode of forecasting has so far shown unprecedented accuracy.
Freshwater resource abuse is a serious issue that is occurring in many countries. Using this system emphasizes the idea that you cannot have just one way of treating aquifers as each are unique and has a variety of differing factors that must be taken into conisderation. These forecasts have already helped raise awareness among citizens of the Andhra Pradesh region of India. This will hopefully prompt government officials to devise or improve current water stewardship practices in India and incite similar studies in other countries.
-Elisabeth Shapiro
Sources:
Banerjee Pallavi, Prasad R.K, and Singh V.S (2008). Forecasting of groundwater level in hard rock region using artificial neural network. Environ Geol.
It's neat how the predictions can be made for such a long period of time. It is probably very convenient for certain regions. I think the people making the predictions should also look into the countries that consume a large amount of water, such as Canada, and discuss how to use less water. Many of us Canadians seem to believe that our water supply is limitless. A study made known throughout such countries could be very benefitial.
ReplyDeleteThis is an amazing step forward in planning for the future of our water resources. It makes me wonder if the same mathematical principles can be put toward other types of water use. It would be very helpful if we could use it for things like pollution.
ReplyDelete