Scalable Methods for Citizens to Measure Spring Discharge - a Canary in a Coal Mine for Changing Hydrological Systems

A canary in a coal mine. Most people are familiar with the story: coal miners bringing caged canaries into new coal seams to utilize their sensitivity to methane and carbon monoxide as an early warning sign of dangerous gas concentrations. As long as the bird kept singing, the miners knew their air supply was safe; a quiet (or dead) canary meant something could be wrong. With regards to water resources, if we aim to wisely steward them, we must first learn to measure them. However, are there any "canaries" that we should place specific emphasis on measuring? We suggest that headwater springs are "canaries" that should receive specific attention. Changes in spring discharge or quality can be a warning sign of unsustainable water extractions, changes to land use, geology, or precipitation patterns. Additionally, springs (in Nepal and elsewhere) are critical water sources, especially during prolonged dry periods (e.g. pre-monsoon in Nepal) without measurable precipitation. To monitor the springs of the world, we need new and scalable measurement methods. Towards this goal, SmartPhones4Water-Nepal (S4W-Nepal) evaluated three different citizen science spring discharge measurement approaches in the spring of 2017: float, salt dilution, and Bernoulli. Linear regressions forced through the origin for scatter plots with reference flows had slopes of 1.05, 1.01, and 1.26 with r-squared values of 0.90, 0.98, and 0.61, for float, salt dilution, and Bernoulli methods, respectively. After selecting the preferred approach (salt dilution), we developed an Open Data Kit (ODK) Collect based application and piloted the approach in the Kathmandu Valley with 20 student volunteers. After one day of training and one week in the field, volunteers performed over 145 discharge measurements ranging from 0.4 to 425 liters per second. Unfortunately, there was no systematic way to evaluate the accuracy of all the measurements performed. However, at the five locations were S4W-Nepal FlowTracker measurements were available, the resulting salt dilution errors (𝜇 = -6.7%, SD = 11.5%) where comparable to our initial evaluation data (𝜇 = 8.2%, SD = 17.2%). Motivated by these promising results, future work should further evaluate the applications of citizen science based salt dilution streamflow measurements to a larger area of Nepal and beyond.