proudly sponsored by
|
archiving of SOWACS proudly sponsored by |
|
[Apologies: my comment placed at the beginning of the "Standardising Sensors made it appear as if I (Bruce) had originated it: it was in fact a post from Tom A Reynolds who put forward the idea. Due to the fact that the mailing list is moderated - to filter junkmail - each and every mail passes through my hands, and I occasionally omit the originator's details. I normally just send posts out to sowacs without editing, but this time will take Richard up on his offer to respond... Bruce Metelerkamp - owner of sowacs - my comments in [brackets] ] Richard Mead writes Bruce and others on SoWaCS, First, it would be nice if the recent individual submitting the calibration comments would sign his/her postings. Please prey-tell, who submitted the query to the SoWaCS group? [As explained above - this was my omission.] Secondly, why is it that 'some' sensor companies think that calibrating their sensors in quartz sand is the ultimate way to calibrate a sensor? A sensor that will inevitably NOT be working in sand? If sand is such an easy medium, it doen't seem to offer much of a range of soil moisture. How about calibrating in various soil types [sand, loam, clay loam etc.], bulk densities, temperatures, salinities, organic matter content etc.? [Ideally, what one wants to know when evaluating a sensor, is the accuracy and range over which it can reliably be used. The companies need to provide an indication of these ranges and limits ...but without having to employ teams of technicians working around the clock... the sensors are expensive enough already, are they not?] Moving right along....some answers/comments to the questions posed: >What should we standardize on for a container? Okay, if we decide on a container, then how would field calibrations correlate? Why can't we stick to the standard gravimetric procedure (oven dry bulk density samples for 48 hrs at 105 C)? I think the location of gravimetric sampling to the sensor monitoring area is more important than the size of a container and measuring the container's weight. The larger the container (albeit the more accurate), the more of a hassle it would be to find a scale to accommodate the weight...soil and water are heavy you know? [This comes back to the opposing arguments of undisturbed vs reconstituted samples. The choice depends heavily on the sensor type and soil type. If the soil structure is such that it cannot be adequately reconstitued in a container - then what use is calibrating that way? To work back to volumetric adequately, preventing changes to the bulk density is critical.] >What spread sheet are we all going to use to file the data? Don't think the type of spread sheet matters, it's the statistical number crunching that should be standardized (regression coefficients, confidence intervals, minimum number of replications etc.) [Is that not what is being inferred by the "type" of spreadsheet? There is a very illuminating paper that investigates the number of sampling points, their spacing and measurement repetitions for neutron probe use. Depending on your stance - the requirements are surprisingly high, yet some indication can be obtained with a surprisingly few measurements. (Unfortunately, I cannot recall the paper, and I no longer hae access to the collection.)] >What about using room temperature? Why just one temperature? How often does room temperature happen outdoors? Why not 10, 20, 30 and 40 C? [What those of us in less warm climates? :(] >Should we repeat process, using three different sensors? Definitely, a minimum number of three sensors each with a minimum number of three readings per increment of soil moisture. [There are those of us - perhaps even the majority - that just want to get on with _using_ the (one? type of) sensor to get some data to work with. Perhaps a case of a litle knowledge being dangerous? But then at least an attempt at assessing soil water content is better than none!] Good topic, let's hammer it out! Comments/curses to my info submitted above are welcomed. Regards, Richard Mead Agrilink International Soil Scientist & Technical Director (209) 324-9324 Phone (209) 324-9325 Fax