Midwestern cities require forecasts of surface nitrate loads to bring
additional treatment processes online or activate alternative water
supplies. Concurrently, networks of nitrate monitoring stations are
being deployed in river basins, co-locating water quality observations
with established stream gauges. Here, we construct a synthetic data set
of stream discharge and nitrate for the Wabash River Basin - one of the
U.S.’s most nutrient polluted basins - using the established Agro-IBIS
model. While real-world observations are limited in space and time,
particularly for nitrate, the synthetic data set allows for sufficiently
long periods to train machine learning models and assess their
performance. Using the synthetic data, we established baseline 1-day
forecasts for surface water nitrate at 12 cities in the basin using
support vector machine regression (SVMR; RMSE 0.48-3.3 mg/L). Next, we
used the SVMRs to evaluate the improvement in forecast performance
associated with deployment of additional sensors. Synthetic data enable
us to quantitatively assess the expected value of an additional nitrate
sensor being deployed, which is, of course, not possible if we are
limited to the present observational network. We identified the optimal
sensor placement to improve forecasts at each city, and the relative
value of sensors at all possible locations. Finally, we assessed the
co-benefit realized by other cities when a sensor is deployed to
optimize a forecast at one city, finding significant positive
externalities in all cases. Ultimately, our study explores the potential
for AI to make short-term predictions and provide an unbiased assessment
of the marginal benefit and co-benefits to an expanded sensor network.
While we use water quantity in the Wabash River Basin as a case study,
this approach could be readily applied to any problem where the future
value of sensors and network design are being evaluated.
People Detection in a Depth Sensor Network via Multi-View CNNs trained on Synthetic Data