The Navajo Nation Department of Water Resources (NNDWR) manages a network of rain cans, automated climate stations, stream gauges and snow monitoring gauges (both Snow Courses and SNOTEL sites). The majority of these stations must be visited on a monthly basis by NNDWR staff to gather data and inspect the gauges for damage. Nationwide budgetary constraints have forced the NNDWR to re-evaluate its gauge network to reduce field costs.
Lynker was hired to reduce the field effort required to maintain the NNDWR gauge network while optimizing the network’s ability to accurately predict rainfall across the entire Navajo Nation. At the project start, there were more than 110 stations that require regular visits across 27,000 sqmi of land, requiring a very large amount of monthly field work to visit, maintain and collect data from all the stations. The project team implemented an ArcGIS model to reduce by 30% the monthly field hours required to visit and monitor the rain gauge network while minimizing the measurement error introduced to the system by the removal of gages.
The impact of an increase in measurement error is best described in terms of the water that is “missed” as a result of the removal of a gauge. The Volumetric Index of the Error of Water (VIEW) is the error in measurement multiplied by the average precipitation at any given grid cell. The average precipitation used was the PRISM 30-year annual average dataset.
With a stringent requirement for field effort reduction, the primary means of network optimization was the removal of tipping bucket rain gauges. A combination of factors led to the creation of optimized network test cases; gauge period of record, clustering with other gauges, rainfall measurement coverage within individual watersheds, and a qualitative evaluation by NNDWR staff of data quality and potential for vandalism at the gauge site.
Automated climate gauging stations were recommended to either be connected to the central NNDWR office through radio or cellular telemetry or were removed altogether to offset the high associated maintenance cost. To improve precipitation prediction, scripted connections were created to existing data sources such as the NOAA-COOP rainfall network so the NNDWR could benefit from easy access to that data.