Assessment of soil water repellency (SWR) was conducted in the decomposed organic floor layer (duff) and
in the mineral soil layer of two Mediterranean pine forests, one in Italy and the other in Spain, by the widely-used water
drop penetration time (WDPT) test and alternative indices derived from infiltration experiments carried out by the
minidisk infiltrometer (MDI). In particular, the repellency index (RI) was calculated as the adjusted ratio between
ethanol and water soil sorptivities whereas the water repellency cessation time (WRCT) and the specifically proposed
modified repellency index (RIm) were derived from the hydrophobic and wettable stages of a single water infiltration
experiment. Time evolution of SWR and vegetation cover influence was also investigated at the Italian site. All indices
unanimously detected severe SWR conditions in the duff of the pine forests. The mineral subsoils in the two forests
showed different wettability and the clay-loam subsoil at Ciavolo forest was hydrophobic even if characterized by organic
matter (OM) content similar to the wettable soil of an adjacent glade. It was therefore assumed that the composition
rather than the total amount of OM influenced SWR. The hydraulic conductivity of the duff differed by a factor of 3.8–
5.8 between the two forested sites thus influencing the vertical extent of SWR. Indeed, the mineral subsoil of Javea
showed wettable or weak hydrophobic conditions probably because leaching of hydrophobic compounds was slowed or
prevented at all. Estimations of SWR according to the different indices were in general agreement even if some discrepancies
were observed. In particular, at low hydrophobicity levels the SWR indices gathered from the MDI tests were able
to signal sub-critical SWR conditions that were not detected by the traditional WDPT index. The WRCT and modified
repellency index RIm yielded SWR estimates in reasonable agreement with those obtained with the more cumbersome RI
test and, therefore, can be proposed as alternative procedures for SWR assessment.
Beerkan infiltration runs could provide an incomplete description of infiltration with reference to either the near steady-state or the transient stages. In particular, the process could still be in the transient stage at the end of the run or some transient infiltration data might be loss. The Wu1 method and the BEST-steady algorithm can be applied to derive soil hydrodynamic parameters even under these circumstances. Therefore, a soil dataset could be developed using two different data analysis methods. The hypothesis that the Wu1 method and BEST-steady yield similar predictions of the soil parameters when they are applied to the same infiltration curve was tested in this investigation. For a sandy-loam soil, BEST-steady yielded higher saturated soil hydraulic conductivity, Ks, microscopic pore radius, λm, and depth of the wetting front at the end of the run, dwf, and lower macroscopic capillary length, λc, as compared with the Wu1 method. Two corresponding means differed by 1.2–1.4 times, depending on the variable, and the differences appeared overall from moderate to relatively appreciable, that is neither too high nor negligible in any circumstance, according to some literature suggestions. Two estimates of Ks were similar (difference by < 25%) when the gravity-driven vertical flow and the lateral capillary components represented the 71–89% of total infiltration. In conclusion, the two methods of data analysis do not generally yield the same predictions of soil hydrodynamic parameters when they are applied to the same infiltration curve. However, it seems possible to establish what are the conditions making the two methods similar.
The extent (determined by the repellency indices RI and RIc) and persistence (determined by the water drop penetration time, WDPT) of soil water repellency (SWR) induced by pines were assessed in vastly different geographic regions. The actual SWR characteristics were estimated in situ in clay loam soil at Ciavolo, Italy (CiF), sandy soil at Culbin, United Kingdom (CuF), silty clay soil at Javea, Spain (JaF), and sandy soil at Sekule, Slovakia (SeF). For Culbin soil, the potential SWR characteristics were also determined after oven-drying at 60°C (CuD). For two of the three pine species considered, strong (Pinus pinaster at CiF) and severe (Pinus sylvestris at CuD and SeF) SWR conditions were observed. Pinus halepensis trees induced slight SWR at JaF site. RI and RIc increased in the order: JaF < CuF < CiF < CuD < SeF, reflecting nearly the same order of WDPT increase. A lognormal distribution fitted well to histograms of RIc data from CuF and JaF, whereas CiF, CuD and SeF had multimodal distributions. RI correlated closely with WDPT, which was used to develop a classification of RI that showed a robust statistical agreement with WDPT classification according to three different versions of Kappa coefficient.
In Mediterranean ecosystems, special attention needs to be paid to forest–water relationships due to water
scarcity. In this context, Adaptive Forest Management (AFM) has the objective to establish how forest resources have to
be managed with regards to the efficient use of water, which needs maintaining healthy soil properties even after
disturbance. The main objective of this investigation was to understand the effect of one of the AFM methods, namely
forest thinning, on soil hydraulic properties. At this aim, soil hydraulic characterization was performed on two
contiguous Mediterranean oak forest plots, one of them thinned to reduce the forest density from 861 to 414 tree per ha.
Three years after the intervention, thinning had not affected soil water permeability of the studied plots. Both ponding
and tension infiltration runs yielded not significantly different saturated, Ks, and unsaturated, K–20, hydraulic conductivity
values at the thinned and control plots. Therefore, thinning had no an adverse effect on vertical water fluxes at the soil
surface. Mean Ks values estimated with the ponded ring infiltrometer were two orders of magnitude higher than K–20
values estimated with the minidisk infiltrometer, revealing probably soil structure with macropores and fractures . The
input of hydrophobic organic matter, as a consequence of the addition of plant residues after the thinning treatment,
resulted in slight differences in terms of both water drop penetration time, WDPT, and the index of water repellency, R,
between thinned and control plots. Soil water repellency only affected unsaturated soil hydraulic conductivity
measurements. Moreover, K–20 values showed a negative correlation with both WDPT and R, whereas Ks values did not,
revealing that the soil hydrophobic behavior has no impact on saturated hydraulic conductivity.
The lateral saturated hydraulic conductivity, Ks,l, is the soil property that mostly governs subsurface flow in
hillslopes. Determinations of Ks,l at the hillslope scale are expected to yield valuable information for interpreting and
modeling hydrological processes since soil heterogeneities are functionally averaged in this case. However, these data
are rare since the experiments are quite difficult and costly. In this investigation, that was carried out in Sardinia (Italy),
large-scale determinations of Ks,l were done in two adjacent hillslopes covered by a Mediterranean maquis and grass, respectively,
with the following objectives: i) to evaluate the effect of land use change on Ks,l, and ii) to compare estimates
of Ks,l obtained under natural and artificial rainfall conditions. Higher Ks,l values were obtained under the maquis than in
the grassed soil since the soil macropore network was better connected in the maquis soil. The lateral conductivity increased
sharply close to the soil surface. The sharp increase of Ks,l started at a larger depth for the maquis soil than the
grassed one. The Ks,l values estimated during artificial rainfall experiments agreed with those obtained during the natural
rainfall periods. For the grassed site, it was possible to detect a stabilization of Ks,l in the upper soil layer, suggesting that
flow transport capacity of the soil pore system did not increase indefinitely. This study highlighted the importance of the
experimental determination of Ks,l at the hillslope scale for subsurface modeling, and also as a benchmark for developing
appropriate sampling methodologies based on near-point estimation of Ks,l.
In bare soils of semi-arid areas, surface crusting is a rather common phenomenon due to the impact of raindrops. Water infiltration measurements under ponding conditions are becoming largely applied techniques for an approximate characterization of crusted soils. In this study, the impact of crusting on soil hydraulic conductivity was assessed in a Mediterranean vineyard (western Sicily, Italy) under conventional tillage. The BEST (Beerkan Estimation of Soil Transfer parameters) algorithm was applied to the infiltration data to obtain the hydraulic conductivity of crusted and uncrusted soils. Soil hydraulic conductivity was found to vary during the year and also spatially (i.e., rows vs. interrows) due to crusting, tillage and vegetation cover. A 55 mm rainfall event resulted in a decrease of the saturated soil hydraulic conductivity, Ks, by a factor of 1.6 in the inter-row areas, due to the formation of a crusted layer at the surface. The same rainfall event did not determine a Ks reduction in the row areas (i.e., Ks decreased by a non-significant factor of 1.05) because the vegetation cover intercepted the raindrops and therefore prevented alteration of the soil surface. The
developed ring insertion methodology on crusted soil, implying pre-moistening through the periphery of the sampled surface, together with the very small insertion depth of the ring (0.01 m), prevented visible fractures. Consequently, Beerkan tests carried out along and between the vine-rows and data analysis by the BEST algorithm allowed to assess crusting-dependent reductions in hydraulic conductivity with extemporaneous measurements alone. The reliability of the tested technique was also confirmed by the results of the numerical simulation of the infiltration process in a crusted soil. Testing the Beerkan infiltration run in other crusted soils and establishing comparisons with other experimental methodologies
appear advisable to increase confidence on the reliability of the method that seems suitable for simple characterization of crusted soils.