| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
USDA-ARS Jornada Experimental Range Las Cruces, NM 88003
jherrick{at}nmsu.edu
My coauthor and I would like to thank Dr. Minasny and Dr. McBratney for their perceptive comments. We concur that by accounting for momentum, their Eq. [8] may more accurately reflect the energy transmitted from the sliding hammer to the penetrometer described in Herrick and Jones (2002). We also agree that their new estimate of resistance (Eq. [9]) remains "an experimental measurement which does not give a unique soil property; rather it is a complex measure which depends on both the penetrometer and the soil condition" (Minasny and McBratney, 2005).
Both statements are supported by an empirical test in which we compared the amount of energy required to push a 2.61-kg shaft 15 cm into the soil using 1-, 2-, and 4-kg hammers. The comparison was replicated at 10 randomly selected locations in a flood-irrigated pasture on the New Mexico State University Experimental Farm, resulting in a randomized complete block design. In theory, the energy required for all three hammers should have been identical. The relative difference in the average values for the three hammers generated by Eq. [8] of Minasny and McBratney (2005) (18.4 ± 4.0, 20.0 ± 3.6, and 25.4 ± 3.3 J) was somewhat smaller than differences in the average values generated by the equation reported in Herrick and Jones (2002) (66.6 ± 14.5, 46.6 ± 8.4, and 42.3 ± 5.5 J) (mean ± SD for 1-, 2-, and 4-kg hammers, respectively). This provides empirical support for the new equation. However, the new equation failed to eliminate the effect of hammer mass, which remained highly significant (F2,18 = 10.3; P = 0.001). The latter finding is likely due to a combination of instrument characteristics (e.g., mass-dependent variability in the elasticity of the collision) and mass-dependent soil–penetrometer interactions. It provides support for their statement that penetrometer measurements are the result of complex interactions between the soil and the instrument.
The complexity of these relationships does not, however, prevent penetrometers from being used to monitor changes in soil structure across time. As can be demonstrated mathematically, statistical comparisons among treatments are independent of the equation used. This is reflected empirically because the same coefficient of variation (CV) applies to both equations for replicate measurements made using the same hammer (21.8, 17.9, and 12.9% for a 1-, 2-, and 4-kg hammer, respectively). Identical CVs also result from an analysis of the raw data (number of strikes required to push the shaft 15 cm into the soil).
In conclusion, penetrometer data should continue to be treated as relative values, and comparisons should be limited to penetrometers of similar mass and dimensions regardless of the equation used. This restriction is similar to that applied to comparisons generated using the more common strain gauge penetrometers: treatment comparisons must be limited to instruments with similar dimensions that are inserted at a consistent rate (Bradford, 1986). Where the primary objective is to monitor change or make comparisons to a reference area, we recommend reporting the number of strikes per depth increment (e.g., Herrick et al., 2002) rather than using an energy or resistance indicator. A strike-based indicator is more easily communicated to land managers (Herrick et al., 2005). By maintaining the original data, reporting the number of strikes also allows subsequent application of different computational approaches.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
