proudly sponsored by
|
archiving of SOWACS proudly sponsored by |
|
Dear Peter, Dear Listmembers, Looking at the data one usually gets from soil moisture sensors and their scattering, maybe shouting is not such a bad idea ;-) But why is that ? The basic problem of soil moisture sensors is the small sampled/measured volume, which gives a much higher measurement scattering, than the variability, that the plants see , since they average over a much larger soil volume with their roots. Additionally, the range of soil moisture, in which plants can grow, maps to only a small fraction of the measurement range of sensors, like 100 percent avaiable field capacity might map to only 15 volume percent of soil water. The range, in which you might want your plants to grow in, is even just a fraction of that. An error of, let s say, one volume-percent in a sensor reading (from measurement errors or from variability doesn t matter) expands to maybe ten or twenty percent of the range that you want your crops to grow in. So the solution is not a yet more sophisticated sensor which does perfect measurements on ten cubiccentimeters of soil, but one that averages (evenly!) over 125 liters of soil. Which is still to be designed. Or you use a large number of cheap sensors for an average. That s why I think Peter is right; first you need the agronomist to interprete any given data, including the confidence interval of that data, collected by sensors or by other means, then you might think about investing in better sensors, not the other way around. Botes, Grove and Oosthuizen ( An Economic Analysis of Irrigation with limited water supplies , Sustainable Irrigation in Areas of Water Scarcity and Drought, conference proccedings, Oxford, UK, 11-12 Sept, 1997) put some numbers to it. Best Regards Martin Schmitz ----------------------------- Martin Schmitz Institut fuer Betriebstechnik FAL Germany schmitz@bt.fal.de Tel.: ++49 (0)531 596 430 -----------------------------