Dickey, Gylan L. :

Hanson, Blaine R

When measuring soil water content . . . Field practices affect neutron moisture meter accuracy Blaine R . Hanson o Gylan L . Dickey installed inside the meter's shielding ; Field practices formeasuring soil each standard count should be made water content with neutron mois - under similar conditions . ) ture meters were assessed . One The amount of hydrogen in the soil ofthe same finding:Usingmeters mainly depends on soil water content and soil organicmatter content . Soil manufacturerwith differentcom - mineralogymay also influencethe ponents ( suchas differenttypes amount of slow neutrons . These factors ofdetector tubes ) or meters of dif - require that a calibrationcurvebe uusedre ferentmanufacturerscan greatly to relate counts of the neutron moist alterperformanceand affectre - meter with soil water content . Calibra - tion curves are developed by sampling sults . Guidelines were developed the soil over a range of water contents formaking standard counts and and obtaining a neutron meter count for soil . formaking counts in the each sample . Widespread use of the neutron meter has resulted The neutron moisture meter has been a in a diversity of practices . Some operators use aluminum access valuable and time - savinginstrument for monitoring soil water content . Initially tubing ; others use steel or plastic . Neu - ed Above , the neutronmoisturemeter is used by scientistsfor research , the neu - tron meters are frequently interchang widely used by consultantsto develop irri - gation schedules . without consideringdifferences in each tron meter is now used by state and fed - eral agencies , private consultants and meter's performance.Also , standard growers to monitor soilwater content count checks may be made under a vari - ety of conditions , such as with the meter for irrigation scheduling . The neutron meter indirectly mea - on the ground surface , in vehicle trunks , Effect of access tubing material sures soil water content by first emitting on top of the instrument's carrying case , Measurementswere made in fine so on . fast neutrons from a radioactive source . and sand , loam , stratified clay loam and clay In response to questions concerning As the neutrons speed through the soil , soils using aluminum , steel and loam use of the neutron meter in measuring they lose energy and are slowed by colli - soil water content , a study was con - plastic ( PVC Schedule125 ) tubing . Tub - sions with hydrogen particles . This re - 1 . ing dimensions are shown in table ducted to determine : sults in a " cloud of slow neutrons in the Calibrationcurves were developed by vicinity of the detector of the neutron ( 1 ) The effect of access tubing mate - sampling the rial on the count rate of neutron meters . soil at 6 - inch depth inter - ss probe . The higher the soilwater content , vals with a Madera sampler . The acce ( 2 ) The effect of interchanging meters . the denser the cloud of slow neutrons . tube then was installed in the hole used Operation of the neutron meter re - ( 3 ) Appropriate counting rates and for soil sampling.Measurementswere quires installing an access tube in thef number of counts per depth.s then made with the neutron meter at 6 - soil . The meter is mounted on top o the ( 4 ) Standardized method for making 5 feet deep . Counting inch intervals to access tube , and the source tube ( which standard count checks . time was 30 seconds.Soilwater con - containsboth the radioactive source that tents , obtained from the samples , were is emitting fast neutrons and the slow - then correlatedwith the counts . neutron detector ) is lowered into the ac - Figure 1shows the counts versus cess tube . The source tube is lowered to depth for the three tubing materials for a particular depth and a count is made . the clay loam ( fig . la ) and the fine sand The operator makes a standard count ( fig . lb ) . For the clay loam soil , the high - periodicallyto , monitor the meter's per - est counts occurredwith aluminum ac - h formance and to identify any problemsis cess tubing and the lowest counts wit with its operation . ( A standard count a plastic tubing . Counts with steel tubing set of counts made with the source tube CALIFORNIA AGRICULTURE , NOVEMBER - DECEMBER1993 29 ( fig . 3b ) . Calibrationcurves for plastic ( b ) Fine sand 4,000 cylinderswere the same for practical @ 12,000 purposes for the four meters . However , P for the water content meter with the smallerneutron source , the calibration 8,000 2,000 curve still differed considerablyfrom the 6,000 curves of the other instruments . 2 4,000 Use caution in interchanging meters if an instrument is repaired or compo - i 2,000 + Aluminum + - Steel 0Plastlc - r Alumlnum - D Steel 0 Plastic nents upgraded . This may change the calibrationcurve . Figure 4 shows cali - 6 12 18 24 30 36 42 48 54 60 6 12 18 24 30 36 42 48 54 80 Depth below groundsurface ( in . ) Depth below ground surface ( in . ) bration curves for two water contentut meters of the same manufacturer , b Fig . 1 . Counts per depthfor aluminum , plastic and steel access tubing in ( a ) a clay loam with considerablydifferent 30 - second soil and ( b ) a fine sand . standard counts , indicating different de - tector characteristics.The calibration curves made in a sandy loam soil show large differencesin the behavior of the were about midway . As the counts in - Interchanging moisture meters two instruments , even though count ra - creased with depth , differencesbetween tios were used . Counts were made in polyethylene counts of the various tubing also in - These results indicate that if the count plastic cylinderswith five neutron water creased . Similarbehavior also occurred ratio is used , meters of the same manu - content meters . Wall thickness of the cyl - for the loam and stratified clay loam facturer with similar components can be inders ranged from 0.45 to 2.95 inches . sites . For the fine sand , however , little interchanged . Meterswith very different Calibrationcurves of counts versus wall differencein counts with access tubing characteristicsshould not be interchanged thicknesswere developed for each water material occurred because of the very unless a correctionfactor is used to adjust contentmeter . Four of the meters were low counting rate in this soil . the calibrationcurve.Moreresearch is of the same manufacturer , each with a Figure 2 shows the calibrationcurves needed , however , before a procedure can source strength of about 50 millicuries : . for the clay loam site for each access tub - be recommended . The fifth meter had a source strength of ing material . For a given water content , about 10 millicuries . a higher count occurred for aluminum Standardized method for counts Figure 3a shows the calibration tubing , an intermediate count for steel Frequent standard counts should be curves for the plastic cylinders.A con - tubing and a relativelylow count for made to monitor the performance of the siderable difference exists in the counts plastic tubing . The slope of the calibra - meter and to calculate the count ratio . among the four meters of the same tion curve was the smallestfor plastic The standard count normally is tmadehe manufacturer . For a wall thickness of tubing and the highest for aluminum with the radiation source locked in 1.28inches , counts ranged from about tubing . Since the slope of the calibration instrumentâ??s shielding . The height of cthee 12,000 per 30 secondsto about 17,000 per curve is the change in counts with meter above the ground or other surfa 30 seconds.Differencesin counts among changein water content , this behavior - such as the carrying case - must be the meters increased as the wall thick - indicates that the count sensitivityto sufficient to prevent any surfacereflec - ness of the cylindersincreased.Rela - changesin water contentis the largest tance interference.The horizontal dis - tively small counts occurred for the with aluminum tubing and the smallest tance between the meter and any body meter with the 10millicurie sources , with plastic tubing . This behavior was of matter high in hydrogen affectingthe which would be expected . Water content consistentat all sites except for fine standard count , such as the operator ror , meters of the same manufacture can sand . the presence of another neutron mete vary in count rate , probably because of The behavior in figures 1and 2 is should alsobe of concern.Meter shield - variability in the source strength and dif - causedby material in the tubing which ing is only 30 to 70 % effective.Escaping feringcharacteristicsin the instrumenfs absorbsneutrons . Data from the Interna - neutrons may be reflected back into the electronics.nt tional Atomic Energy Agency ( Vienna ) meter and increasethe standard count . If a cou ratio ( actualcount in the show that chlorinecan absorbmanyalumi - and distances at which The heights cylinder divided by a standard count ) is more neutrons than either iron or interferenceoccurred were determined used , differences among meters of the num . Iron is a distant second and alumi - by first positioning the instrument at same manufacturer become negligible num has little ability to absorb neutrons . various heights aboveground and then Thus , the chlorinein plastic ( PVC ) tub - taking 30 counts , each 30 seconds long . ing causesthe smallest countsbecause of 16,000 - The instrument was then positioned at neutron absorption . Counts in steel tub - h 4 14,000 various horizontal distances from a 5 - ing are relativelyhigher compared wit g 12,000 gallon container of water and from an - plasticbecause iron absorbs far fewer 8 10,000 other neutron water contentmeter . neutrons than does PVC . Counts arese 8,000 Results show that for heights greater highest in aluminum tubing becau 6,000 than 19 inches , the ground surfacehad relatively little neutron absorption oc - 4,000 little effect on the standard count . Mea - curs in this material.Absorption of neu - surements made at heights greater than trons also changesthe slope of the cali - 0 I I I I 19inches showed nstatisticallyinsignifi - t bration curve ( fig . 2 ) . Using aluminum 0 0.1 0.2 0.3 0.4 cant differencesfrom the measureme Soil water content ( In . I In . ) access tubing would allow a neutron made at 19inches ; counts made at meter to sense a smaller change in soil heights less than 19 inches showed sta - Fig . 2 . Calibrationcurves for the three dif - water content because neutron absorp - tistically significant differencesfrom that ferent access tubings in a clay loam soil . tion by aluminum is very small . CALIFORNIA AGRICULTURE , VOLUME 47 , NUMBER 6 30 was 5.2 % . The error was 3.3 % for a + - CK - D CD u C384 + C551 + standard count value based on the 3.1 % for five - count average and was the ten - count average . Differencesin errors were statisticallysignificant ( 90 % confidencelevel ) between the one - count value and the five - count average.Differ - ences between the five - and ten - count averagewere not statisticallysignificant . This indicates that a standard ucountlt 0.57 0.82 1.04 1.28 0.57 0.82 1.04 1.28 based on one count only could res in Cylinder wall thickness ( in . ) Cylinderwall thickness ( in . ) relativelylarge errors and that a ten - Fig . 3 . Calibrationcurves for five neutron meters between wall thicknessand ( a ) counts count averagewas no better than a five - and ( b ) count ratio . count average . However , a five - count average did sigruficantlyreduce error and is the recommended procedure . made at the 19 - inchheight . The recom - counts and the count ratio calculated mended minimum horizontal distance from 30 counts . ) Four different calibra - Summary tion curves of water contentversus count between the probe and any material that ( 1 ) Access tubing material substan - tron could affect the standard count is about ratio were used to determine potential tially affects the counting rate of neu errors in calculatingwater content . 25 inches . No statisticaldifferencesex - water contentmeters and their sensitiv - isted for counts made at distances Figure 5 shows how the number of ity count measurements needed to obtain in detecting a change in soil water greater than 25 inches compared with the count at 25 inches ; counts at dis - errors of content.Highest rates were in aluminum 1and 2 % at a confidencelevel tances less than 25 incheswere statisti - tubing and the lowest in plastic tubing . of 90 % varies accordingto total count . cally different . However , the minimum The counting rate was less sensitive to The data sets show that as total count distancebetween a second meter and the changesin increases , fewer count measurements soil water content for steel and plastic tubing . meter used for the standard count are needed . Also , fewer count measure - ( 2 ) Neutron water content meters of should be at least 16feet . ments are needed for the larger error . For counts greater than 25,000 to 30,000 , the same manufacturer and model pwithe Countingtime , counts per depth a one - count measurement per depth is similar components ( such as same ty adequate for a of detector tubes ) can have slightly dif - Both the most appropriate counting 1 % error . One measure - ferent performances . These differences time and the number of count measure - ment is needed for counts of 6,000 to can be partially overcomeby using the ments per depth were investigated by 7,000 with a 2 % error . A one - measure - count ratio instead of actual counts . making 30 counts in each plastic cylin - ment sample will result in an error ofthan However , meters of the same manufac - der . Count times of 15,30 and less than 2 % for total counts greater 60 sec - turer with different components ( suchas 6,000 to 7,000 . The counting time per se onds were used . Counts were made with different types of detector tubes ) or did not affect the sample two neutron probes with respective30 - size , except as meters of different manufacturers canount second standard counts of 10,382 wandas it affectedthe total count . Thus , 15 - sec - gh differ greatly in performance.The c 29,808 . Statistical sampling theory ond counts with the meter with the hi ratio may not compensatefor those dif - used to calculate the number of count standard count gave results similar to measurements needed for errors of that of the 30 - second counts with the 1 ferences . ( 3 ) Standard counts should be made other meter . and 2 % at a confidencelevel of 90 % . at least 19inches above any surface and Errors in water content calculations These results then were used to calculate the potential maximum count ratio er - at least 25 incheshorizontally from any due to counting errors depend on errors material that could affect the count rate . in actual ror , affected by both the actual count soil count and in the standard Any nearby neutron meter should be at error and the standard count error , rand count . The potential maximum water the potential maximum error in wate least content errors were assessedfor a stan - 16feet away . content calculations . ( The count ratio is dard count value based on one count ( 4 ) The value of the standard count the actual count in the soil divided by only , an average of five counts and an used to calculate the count ratio should the average standard count . The count average of ten counts . If only one count be an average of five standard counts . ratio error is the differencebetween the was used , the averagewater content er - ( 5 ) One count measurement per count ratio calculatedfrom one ror ( based on the four calibrationcurves ) depth should be sufficient for total or two counts greater than 6,000 to 7,000 . Counting time is important only as it af - - Stn count = 12,837 o 1 % error + 2 % error For neutron watert - fects the total count . - D Stn count = 26,573 contentmeters with a very high coun ing rate , a 15 - secondcount time may be appropriate . Longer counting times may be needed for meters with small count rates . B . R . Hanson is Irrigation and Drainage H0.40.5 0C . 6ount0r.a7tio 0.8 0.9 1 sr 1 Specialist , Department of Land , Air and 0 10 20 30 40 60 50 0.3 Total counts ( thousands ) WaferResources , UC Davis , and G . L . Dickey is Agriculfural Engineer , Technol - Fig . 4 . Calibrationcurve between two 5 . Number of count measurements Fig . ogy Information System Division , Soil Con - versus total counts needed for counting of same manufacturer but with dif - probes of 1 and 2 % . servation Service , Fort Collins , Colorado . errors ferent components . CALIFORNIA A ( ; RICULTURE , NOVEMBER - DECEMBER1993 31