Comparing Citizen Science-Based Rainfall Measurement from a Local Rain Gauge with a Standard Gauge in the Kathmandu Valley

Abstract:

Citizen science - enabling ordinary people in scientific research - is a promising approach of data collection in a developing country like. Citizen science-based rainfall measurement can help generate and monitor a wide range of rainfall data and in turn, support wise water management. Over the past few years, Smartphones For Water Nepal (S4W-Nepal), a young researcher-led water monitoring network, is collaborating with citizen scientists to generate Spatio-temporal precipitation measurements in Nepal using low-cost (< 1 USD) S4W rain gauges, constructed from repurposed soda bottles. To standardize and evaluate the accuracy of S4W gauge or any other ordinary gauge, it is crucial to compare it with a standard gauge. This study compares 24 hours of cumulative rainfall of S4W rain gauge with the standard rain gauge maintained by the Department of Hydrology and Meteorology, Nepal (DHM) of fourteen co-located stations in the Kathmandu valley during the monsoon (June to September) 2019. The nearest station location of S4W and DHM rain gauges were selected within a range of 1-3 kilometers, and statistical parameters were calculated. For the analysis, only the overlapped rainfall data of two datasets (i.e., from S4W and DHM rain gauge) were considered. The results showed that the correlation coefficient between seven stations was strong and positive in Bhaktapur (r=0.82), Jitpurphedi (r=0.88), Khumaltar (r=0.92), Nangkhel (r=0.85), Panipokhari (r=0.90), Kathmandu (r=0.75) and Nagarkot (r=0.78) whereas Chapagaun (r=0.47) , Changunarayan (r=0.46), Tikathali (r=0.63), Sundarijal Alopat (r=0.64) and Sundarijal Mulkharka (r=0.54) have positive and moderate correlation and Naikap (r=0.26) and Sakhu (r=0.10) have weak and positive correlation. The t-test shows that the rainfall data of S4W and DHM rain gauge in all stations are statistically significant at 0.05 significance level except Sakhu station. The relative bias of Bhaktapur, Kathmandu, Chapagaun, Khumaltar, Naikap, Nagarkot, Nangkhel, Sundarijal Mulkharka, and Tikathali stations were less than ±20, whereas, remaining seven stations showed more than ±20. Furthermore, the rainfall data generated by citizen scientists were found to be reliable and can further help to fulfill the data gaps in data constraint countries like Nepal.