A rain gauge (pluviometer) is an instrument used to measure rainfall in a certain period of time and location. Rain gauges can be divided into two groups: non-recording and recording types. In the latter, three more subtypes exist and are known as weighing-bucket, tipping-bucket, and floating (or natural siphon) type rain gauge.
Here is shown the comparison of a tipping-bucket rain gauge (HOBO brand) with the pluviometer used in SENAMHI (Hellman type, see Andrade et al. for more information). The tipping-bucket- (hobo) and hellman- type (conventional) were evaluated at daily scale in the Quincemil (Cusco) station area (latidude: 13°13'1''S, longitude: 70°45'1''W, elevation: 651 msal).
Raw output data of hobo has a typical format (Figure 1) where the more important variables are: Date Time and Event, which corresponds the time (Date Time) in which an event of rain happened and is registered as a pulse (Event).
Figure 1. Typycal format of a HOBO data logger
To convert this rain pulse (count) to rain (mm), the following steps are needed:
-
Each pulse has a value of 0.2 mm (0.254 in)
-
Events that occur in less than 1 second are not valid. If two events consecutively happened in 1 second, only the first one is preserved, i.e is set to 0 (DOUBLE FALSE)
-
Events that correspond to the time of installation and unloading of data should be set to 0 (INSTALLATION and DATA DOWNLOADING)
Taking into consideration the previous steps, it is possible to have a time series of rainfall at a sub-daily scale (seconds) that can be accumulated at higher scales. Comparing the processed hobo and convential is thus easy to perform (see example.R). A simple plot is carry out in the Figure 2.
Figure 2. Time series of daily rainfall for hobo and conventional data
Is easy to see that hobo match really good with conventional for high and low values in the period time. The mean bias is around -1.8 mm and the spearman correlation has a value of 0.97. All this lets us know that hobo works great! However, more research has to be done!
I really thanks to Boris Ochoa Tochachi who provides me the original .xls file with several considerations (mentioned in the Process section).