From: Michael O'Kane (mao320) To: BRUCE@mailgate.icfrnet.unp.ac.za Date: Saturday, April 27, 1996 8:20 am Subject: Soil Suction and Soil Water Content Bruce: I just sent you a message requesting that I be put on your discussion group. I read through some of the comments thoughts etc. that you have received and put on the disussion group. I would be interested in your comments on the following. One of the comments that kept on coming up was the need for insitu calibration in order to obtain qualitative measurements. I could not agree more. I often think that the worst thing that could have happenned with the TDR was the Topp et al. (1980) equation. Now there are people all over the world simply installing these sensors, using the equation, and obtaining a volumetric water content. Very dangerous, since the readings are only qualitative. This is OK for some applications (although I firmly believe that most people want a quantitative measurement), however more often than not the "numbers" are reported as a quantitative measurement and I would wager that the users also start tending to believe that their numbers are qantitative. For the people who do field calibrate, there are still problems. Among others things, the water content sensor reading is dependent on instu density, (i.e. two samples prepared at the same gravimetric water content but with a different density, will have a different volumetric water content). It is very common, if not the rule, that insitu density changes with depth, location, etc. Therefore, in order properly field calibrate a sensor (i.e. a range of volumetric water contents over the range of the sensor's output) one should obtain gravimetric and bulk density mesaurements over a range of water contents for each sensor installed at each depth. This is never done for the simple matter it is not practical nor possible. My point is that even the users that "field calibrate" their w/c sensors are not really doing so since they are taking a single sample (or an everage of three or four) and then comparing that to the sensor outpout and the calibration curve/equation. THIS DOES NOT PROVIDE AN ADEQUATE MEANS OF CALIBRATING A SENSOR. In short, I feel that while it is possible to field calibrate a water content sensor, very few people are doing it properly, and it is VERY impracticle. This brings me to the main reason that I sent this e-mail. One of the people in the discussion group answered the question: Does the CS615 measure soil water potential. The reply was (paraphrasing): Yes, it can if the w/c readings are reference the the soil water characteristic curve (SWCC). THIS IS FUNDAMENTALLY INCORRECT! [Bruce comments] I think Mike meant to say "..are referenceD TO the soil water characteristic curve.."[Bruce] Soil suction is the key insitu measurement for characterizing soil water storage and movement. When considering total suction it must be understood that it is comprised of two components, matric suction and osmotic suction. Total suction corresponds to the free energy of the soil water, while the matric and osmotic suctions are components of the free energy. Essentially, matric suction is the difference in pressure across the air water interface (Ua - Uw), where Ua = pore air pressure and Uw = pore water pressure. Matric suction is commonly associated with the capillary phenomenon arising from the surface tension of water. Field matric suction values can be thought of as a measurement of the negative water pressure within the soil cover system. These measurements are important since matric suction describes the stress state of the soil cover system. In other words, a change in the degree of saturation or water content of the soil cover system is caused by a change in matric suction within the soil cover. Field measurements of matric suction describe the storage and movement of soil moisture and not the other way around. In addition, field matric suction values provide a means of verifying field water content measurements. Finally, since matric suction describes the stress state, it can be used to determine instu hydraulic gradients. This is very important in the field since a very large percentage of natural and engineered soil systems are layered. Therefore, water content measurements will not indicate the direction and rate of moisture movement. Only the stress state measurement will give you the hydraulic gradient. Measurement of matric suction in the unsaturated zone is as important as measuring hydraulic heads in a saturated groudwater system. A matric suction sensor can be thought of as the "piezometer" of the unsaturated zone. [Bruce comments] I agree with your arguments re soil water content and suction, and although fundamentally correct, because of the small matter of practical difficulty (and sometimes) ingnorance, these principles are often not applied.[Bruce] The statement that one can use the SWCC to get matric suction values ( -ve water potential) from w/c measurements is incorrect! One should only go the other way (i.e use matric suction measurements to get w/c measurements based on the SWCC). This brings me to another point: Measurement of the SWCC of each soil within a soil profile is paramount for properly characterizing the profile. One can (and always does) measure the SWCC's in the lab and then use the lab SWCC to characterize the field soils. Care should be taken to ensure that the tests are conducted such that the field stress state and soil structure are represented in the laboratory as much as possible. It is also imperative to verify these measurements in the field. Stress history and initial moulding water content have the most influence on soil structure (and aggregation) which in turn dominates the nature of the soil water characteristic curve of fine grained soils. Specimens of a particular soil will exhibit dissimilar soil water characteristics if they are prepared at different moulding water contents and possess a different stress history, in spite of having the same texture and mineralogy. The effect of weathering (chemical and physical) on the soil water characteristic curve and the saturated hydraulic conductivity should also be included as part of the characterization. [Bruce comments] Can you explain your term "initial moulding water content" a little more? Is this the water content when the undisturbed core is taken?[Bruce] The only practical way to get an insitu SWCC (to be used for coupled heat and mass transfer, soil-atmosphere modelling) is to measure matric suction AND water content at the same time. Our research group installs both matric suction and water content sensors for every field project in order to obtain field SWCC's for each soil under investigation. With respect to insitu matric suction sensors, we have experince with tensiometers, gypsum blocks, and thermal conductivity sensors (indirectly measures matric suction: 1) Agwatronics, 2) Campbell's new 229 sensor, 3) a sensor we obtained through a joint research program with a Univeristy in mainland China, and 4) we are developing our own matric suction sensor through an industry funded (7 companies) and matching federal funding research program. [Bruce comments] As long as we are aware of the potential for error, and have at the outset attempted to correct for SWCC in differing soil layers, then there is some hope. And your illuminating email will go a long way to doing this.[Bruce] We feel that the thermal conductivity sensors have good promise since the physics of the sensor is correct and problems are more often than not a function of poor sensor quality (i.e. the Agwatronics). I just purchased a number of Campbel's 229 suction sensors and we are developing a laboratory calibration curve for each sensor (i.e sensor output versus applied matric suction using the axis translation technique). We feel it is important to develop a laboratory calibration curve for each sensor. We DO NOT use the manufacturer's calibration curve. With respect to water content sensors, we have experience with neutron probe, capacitance type sensors (EnviroScan), TDR (Trace and Campbell's system), and the new CS615. They all have their better or worse applications and depending on budget, insitu soils, location etc. can all be used adequately. [Bruce comments] Are the four sensor types you mention ALL thermal conductivity sensors? With thermal conductivity sensors does one apply a pulse of heat at a known energy rate, and then measure the rate of cooling? If so, then the measuremnt of temperature at points far removed from the datalogger and the use of multiplexers both pose problems. How do you get around these? Is there another name for Campbell's 229 sensor? Could you briefly explain the axis translation technique? (Is it regress y against x during the calibration, and then use x as the independant thereafter)? Are you saying that all the four thermal conductivity (matric suction) sensors that you are testing still do not meet the requirements: >>>THE ONLY THING THAT REMAINS, IS TO DEVELOP A SENSOR(S) THAT ACCURATELY, RELIABLY, AND ECONOMICALLY MEASURES MATRIC SUCTION IN THE FIELD. [Bruce] I guess I've babbled a here. Sorry, but I feel it is very important for people other than in research (i.e. consulting and industry) realize that a stress state is indepent of the material. Therefore, a matric suction sensor does not need to be field calibrated! This is MAJOR advantage over a water content sensor, with out even accounting for the fact water content does not tell you about stress state and direction of water flow in a layered soil system. THE ONLY THING THAT REMAINS, IS TO DEVELOP A SENSOR(S) THAT ACCURATELY, RELIABLY, AND ECONOMICALLY MEASURES MATRIC SUCTION IN THE FIELD. Regards, Mike O'Kane Research Engineer Unsaturated Soils Research Group (USG) Department of Civil Engineering University of Saskatchewan Saskatoon, SK. Canada ========================================================== Bruce Metelerkamp SOIL WATER RESEARCH OFFICER Institute for Commercial Forestry Research, University of Natal, PO Box 100281 Scottsville, ZA3209 Rep. of South Africa Voice:27 331 62314 E-mail: bruce@icfr.unp.ac.za FAX:27 331 68905 URL http://www.icfrnet.unp.ac.za/~metele /SoWaCS.html /RR.html Maintainer of the SoWaCS list.