Comment on ‘reasonable assumptions’

Many of the problems in this course, and in the real world, do not have nice clean questions, data, or answers. As a result, when working on geomorphological type problems you need to constantly be asking yourself the question ‘are my calculations and data sufficient to get an answer at the accuracy I need?’ For many problems you can make reasonable assumptions about data that you don’t have ‘good’ numbers for. As an example, to find the area of a drainage basin you might find it from the USGS data base, or you might carefully planimeter, or use the ‘surveyor’s shoelace method’ to get area from a map (or a google image), or you might just multiply the approximate length of the basin/river by an approximate width. In the first case you are getting somebody else’s fairly accurate data (with unknown errors), in the second case you also get fairly accurate data (with better known errors) and in the final case an estimate with lots of errors. Note none of these are ‘correct’, just somewhat better. The choice of what you use is based on how accurately you need the answer, and that is a judgment call.

Using the correct amount/accuracy of data and work is a real art, and you can only learn by doing. Unlike textbook questions, real world questions never have clean accurate input data. A typical geomorphologic question might need to know ‘how much does it rain in the Laramie area?’ A seemingly innocent question, and a Google search gives the nice answer of 11.5 inches per year, produced by a government agency. However, measuring rainfall is very difficult. Rain gauges sitting only a few meters apart are often vary by 50%. There are never enough gauges to give dense coverage. The rainfall is hugely variable, on virtually all time and space scales. In the Laramie Basin, a lot of rain evaporates before it hits the ground, or almost immediately afterwards. That nice number of 11.5 inches per year is useful, especially to compare to other areas, but is hopeless for most real geomorphic calculations. A particular problem with this number is that it is an average of measured rainfall over a modest length of record, BUT climate is changing. As a result this number could be useful to calculate the rain that fell 25 years ago, but does it really tell us about the current or future rainfall? And I could go on and on. Thus we are always having to ask whether our input data is reasonable, and since we don’t know the answers we are always making assumptions that we have to review in light of the answers.

Learning how to get data to do real world work requires a lot of understanding. It is a learnable art to make assumptions based on our full range of knowledge and experience.