Improving efficiency and repeatability of lake volume estimates using Python

With increasing population and water use demands in Texas, accurate estimates of lake volumes is a critical part of planning for future water supply needs. Lakes are large and surveying them is expensive in terms of labor, time and cost. High spatial resolution surveys are prohibitive to conduct, hence lake are usually surveyed along widely spaced survey lines. While this choice reduces the time spent in field data collection, it increases the time required for post processing significantly. Standard spatial interpolation techniques available in commercial software are not well suited to this problem and a custom procedure was developed using in-house Fortran software. This procedure involved difficult to repeat manual manipulation of data in graphical user interfaces, visual interpretation of data and a laborious manually guided interpolation process. Repeatibility is important since volume differences derived from multiple surveys of individual reservoirs provides estimates of capacity loss over time due to sedimentation. Through python scripts that make use of spatial algorithms and GIS routines available within various Python scientific modules, we first streamlined our original procedure and then replaced it completely with a new pure python implementation. In this paper, we compare the original procedure, the streamlined procedure and our new pure python implementation with regard to automation, efficiency and repeatability of our lake volumetric estimates. Applying these techniques to Lake Texana in Texas, we show that the new pure python implementation reduces data post processing time from approximately 90 man hours to 8 man hours while improving repeatability and maintaining accuracy.