Comment on soil-2021-118

This paper presents a meta-analysis of some papers that deal with the effects of a change from conventional agricultural tillage to some form of conservation tillage (reduced tillage or no tillage). From a literature search, 59 studies were selected following a screening process to include only those studies that provided key data on Ksat, the soil saturated hydraulic conductivity. This is taken by Liao et al. to be the same as steady-state infiltrability measured across the soil surface, though technically this is not the same thing at all. One refers to flow through a saturated porous medium, the other the imbibition of water from free water above the soil to pore water beneath the soil surface. In this case there are interface issues such as surface tension, surface crust and seal effects, the influence of litter, mulch, and other factors. I think that all of this could usefully be clarified in the present ms.

This paper presents a meta-analysis of some papers that deal with the effects of a change from conventional agricultural tillage to some form of conservation tillage (reduced tillage or no tillage). From a literature search, 59 studies were selected following a screening process to include only those studies that provided key data on Ksat, the soil saturated hydraulic conductivity. This is taken by Liao et al. to be the same as steady-state infiltrability measured across the soil surface, though technically this is not the same thing at all. One refers to flow through a saturated porous medium, the other the imbibition of water from free water above the soil to pore water beneath the soil surface. In this case there are interface issues such as surface tension, surface crust and seal effects, the influence of litter, mulch, and other factors. I think that all of this could usefully be clarified in the present ms.
The paper seems to me to neglect some important issues that bear on the interpretation of the published studies. A serious issue for me is that there is no assessment of the quality of the data in the 59 studies. The authors tacitly accept all of the soil Ksat measurements as being valid and reliable measures of Ksat and suitable for their assessment of Ksat differences between forms of tillage.
I do not think that this is a defensible position. It is widely-known, for instance, that the dimensions of the area of volume of soil tested influence the results of many Ksat (or steady infiltration rate) measurements. Thus, if ring or cylinder infiltrometers are used to estimate Ksat from ponded conditions (whether single or double cylinder), the area of soil enclosed within the cylinders (expressed usually by the ring diameter) influences the result obtained. This makes intuitive sense, since a small cylinder might be underlain by a buried stone, so reducing the apparent Ksat, or by a large root macropore, so increasing the apparent Ksat. As the cylinder diameter is increased, the relative effect of such occurrences is reduced. Of course, the choice of an appropriate size of cylinder depends on the properties of the soil being tested. However, the point is that in this paper, Liao et al. simply accept all the results (not mentioning the cylinder diameter used) as valid and meaningful measurements. Unlike, for example, chemical properties, which with care can be measured precisely and unambiguously, the hydraulic properties of soils exhibit a complex dependency on the method and scale of measurement. Some authors have suggested that the measurements need to address something akin to the 'representative elementary volume' concept, adjusted to relate to the scale over which field conditions modulate Ksat.
I mention a few studies here that the authors might find helpful: Another concern that I have with this paper is the choice of test methods for the measurement of Ksat. The authors refer to the effects of drop impact on soil surfaces (e.g. lines 64-65, line 67) via the resulting sealing and crusting effects. Yet none of the measurement methods in their literature survey includes rainfall simulation on field plots, or the study of the response of 'natural plots' (those exposed to real rainfall). It is unclear why such effects should be excluded from analysis. It seems to me to be possible (perhaps probable) that by employing only static water, with no droplet impact, the methods used may well have over-estimated Ksat by excluding dynamic sealing and crusting effects. Likewise, intense rain can drive air into soil pores, thereby reducing infiltrability considerably. Field soils have some residual air content ('field saturation') which is not the same as the 'laboratory saturation' achieved by bottom-up wetting.
There are multiple published studies of conventional versus conservation tillage that do indeed employ rainfall simulation. I have listed a couple of instances below.
A paper that reports rainfall simulation results, including hydraulic conductivity, in a study comparing traditional and conservation tillage is: Packer, I., Hamilton, G., & Koen, T. (1992). Runoff, soil loss and soil physical property changes of light textured surface soils from long term tillage treatments. Soil Research,30(5), 789-806. doi:https://doi.org/10.1071/SR9920789 It is not clear to me why papers such as this were not discovered in the literature survey by Liao et al. If they did not locate this paper, there may be many more that were also not located. It might take a wider choice of search terms than was adopted by Liao et al. to find relevant papers. Overall, therefore, my principal concerns with this paper are the following: (1) the acceptance of published data without evaluation of effects such as the scale of measurement, even when considering just one of the methods, viz., cylinder infiltrometry, and with the lack of reference to studies that employ rainfall simulation methods. The latter have many attendant issues, but at least may capture some of the effects of surface bombardment by water drops, and the development of air entrapment, seals, crusts, etc.
(2) In turn these issues suggest that the authors may need to cast their literature searching net somewhat wider than they appear to have done, as there is a considerable relevant literature, and thorough searching is a cornerstone of thorough meta-analysis work.