Design and set up of a network of disdrometers
Rain keeps still as one of the most difficult meteorological variables to measure nowadays. There are several ways of performing point surface measurements. For some applications precision obtained using manual techniques is more than reasonable. One of the problems with rain is its inherent spatial variability. One can have very reliable measurements at a single point but when they are compared with a close location, important differences are found for the same day. These differences decrease when averaged in time, since they tend to compensate, but sometimes this is not valid for certain applications.Sometimes it is necessary to have daily precipitation fields and, unless a huge number of point measurements are taken, other techniques need to be used. One of these techniques are RADAR and Satellite measurements. With some obvious differences, these techniques estimate the rain curtain falling or not falling to ground from the echo of an electromagnetic wave reflected on the falling drops. Since this is not a direct measurement and considering that sometimes not all these drops fall to ground, some kind of calibration is needed. Two main factors affect he properties of the echo that a curtain of rain produces when a radio beam hits it. One is the size of the droplets, other is the velocity they have. Information on these two factors when precipitation reaches the ground is very valuable for RADAR and satellite rain detection calibration.
The objective of this project was to have 1 minute samples of size and velocity of precipitation particles at 8 close locations with redundant information. All sites were in an area of less than 8 km square.
For this project we used OTT Parsivel disdrometer due to their well known performance. Version 1 of this disdrometer was not specially designed for remote operation, so we had to work hard on the solar power unit in order to guarantee reliability. For datalogging and transmission we went for our own design. It was risky but of the shelf solutions were out of budget or unable to handle the big amount of data. We had to transmit a matrix of wind velocity and size of droplets of more than 1064 cells every minute and coming from 16 different disdrometers. The data was transmitted using GPRS network.
A central server located hundred of km away was listening the remote sites and preparing the data for further post-processing and analysis.
A preventive and corrective quality assurance program with regular visits to the sites was planned for minimizing data loss.
The main reason of success with this project was the design of an intense maintenance program. Disdrometers windows get dirty with dust, animals like to scratch their backs on the structures and spiders use disdrometers arms to build their webs. Besides these minor problems, this project went really smoothly. Some international publications and a Ph Thesis came out of this campaign.