Interactive comment on “ Land-use change effects on soil quality in Montilla-Moriles DO , Southern Spain ”

This manuscript deals with land use change effects on soil organic carbon and nitrogen in a semi-arid Mediterranean agricultural area in Southern Spain, between 1965 and 2011 (46 years). Results show that land use change affected negatively to soil organic carbon and nitrogen contents. In contrast, soil quality (measured as the stratification ratio) was significantly increased. This research is interesting for the scientific community and falls into the scope of SE. However, some revision is necessary before the manuscript is ready for final publication. First of all, the title does not inform adequately of the research carried out and the important questions in the manuscript. I suggest changing the title to: “Impacts of land use change in soil carbon and nitrogen in a Mediterranean agricultural area (Southern Spain)”. In my opinion, results here are widely interesting. So, discussion of results should not stick to a particular area, although the experiments were carried out at farm scale. Results must be put in a C47


Introduction
The soils play a crucial role in influencing the carbon (C) content of the atmosphere because they can either emit large quantities of CO 2 or on the contrary they can act as a store for C (Smith et al., 2000).Agriculture and forestry can achieve this effect through photosynthesis and the C incorporation into carbohydrates (Gonz ález-S ánchez et al., 2012).Crops capture CO 2 from the atmosphere during photosynthesis by converting C forms associated with soil organic matter (SOM) for microbial decomposition processes (Johnson et al., 2007).Soil management is one of the best tools for climate change mitigation and adaptation (Lal et al., 2011).Several Figures

Back Close
Full authors have proposed introducing soil management techniques that combine a restriction on tillage (Corral-Fern ández et al., 2013) and the addition of organic residues (Lozano-García et al., 2011;Lozano-García and Parras-Alc ántara, 2013) to improving soil properties and diminishing atmospheric CO 2 concentrations.Carbon sequestration is defined as any increase in the C content of soils after a change in land management (Powlson et al., 2011) and is one of the most important ecosystem services because of its role in climate regulation (IPCC, 2007).Intensive and conventional tillage (CT) may cause enormous losses of soil organic carbon (SOC), thus inducing an increase in soil erosion and a breakage of soil structure (Melero et al., 2009).
Land use change (LUC) is considered the second greatest cause of C emissions after fuel consumption (Watson et al., 2000).LUC has contributed to soil degradation and soil loss, leading to a decrease in soil C storage worldwide (Eaton et al., 2008), and even more intensely in the Mediterranean areas during the last few decades (Cerd à et al., 2010).Long-term experimental studies have confirmed that soil organic carbon (SOC) is highly sensitive to LUC (Smith, 2008).Thus, even a relatively small increase or decrease in soil carbon content due to changes in land use or management practices, may result in a significant net exchange of C between the soil C pool and the atmosphere (Houghton, 2003).Recently, it has been shown that soil erosion by water and/or tillage has a significant impact on this large pool of SOC (Lal, 2003;Van Oost et al., 2005;Van Hemelryck et al., 2011).
Regional-scale information about C stocks and the relationship between C reservoirs and edaphic factors could be relevant to determine LUC that is of interest in evaluating gains and losses of SOC (Novara et al., 2012).In Spanish soils, climate, use and management are highly influential in the C variability, mainly in the dry Mediterranean climate, characterized by a low SOM content (∼ 10 g kg −1 ), which is a weak and degradable structure (Acosta-Martínez et al., 2003).
The soil C:N ratio is a soil fertility indicator due to the close relationship between SOC and total nitrogen (TN).The soil C:N ratio is often influenced by many factors such as Introduction

Conclusions References
Tables Figures

Back Close
Full climate (Miller et al., 2004), soil conditions (Ou édraogo et al., 2006;Yamashita et al., 2006), vegetation types (Diekow et al., 2005;Puget and Lal, 2005), and agricultural management practices (C ôt é et al., 2000).The concept of using the stratification ratio (SR) as a soil quality indicator is based on the influence of the SOC surface level in erosion control, water infiltration and nutrient conservation (Franzluebbers, 2002).High SR of SOC and TN pools reflect relatively undisturbed soil with high soil quality of the surface layer.The increase of SR can be related to rate and amount of SOC sequestration (Franzluebbers, 2002).
Soil depth has a decisive influence on SOC stocks (Gr üneberg et al., 2010).Some authors have evaluated the SOC content in soil surface (restricted to the upper 15-30 or 50 cm), and few studies have included a deeper section of soil cover (Conant and Paustian, 2002), although vertical processes have a significant impact on SOC variability (VandenBygaart, 2006).Sombrero and Benito (2010) noted that to evaluate and compare SOC storage complete profile is necessary.According to Lorenz and Lal (2005) in temperate climates large amounts of SOC may be stored in subsoil horizons below 30 cm deep.This is essential in LUC because SOC can be transported to a deeper soil horizon, contributing to the subsoil C storage (Lorenz and Lal, 2005).Vertical distribution is one of the features of the organic C store that is not clearly understood together with the relationships with climate and vegetation (Jobb ágy and Jackson, 2000).In the last few decades, a significant number of arable crops (AC) have LUC to olive grove cultivations (OG) or vineyards (V) in Montilla-Moriles Denomination of Origin (DO).This LUC has been motivated by subsidies and better olive oil and wine prices.
Very few reports have compared the effect of LUC from AC to OG and V on SOC and TN storages even on soil quality for long-term in soil entire profiles.In this context, the objectives of this work are (i) to determine the SOC content in the soil; (ii) to study its vertical distribution of SOC (entire soil profile by horizons, non-section control); (iii) to analyze the accumulation and SR of SOC, TN and C:N ratio in Calcic-Chromic Luvisols (LVcc/cr) (FAO, 2006) in AC affected by LUC for long-term in conventional tillage.Introduction

Conclusions References
Tables Figures

Back Close
Full 2 Material and methods

Site description and experimental design
The study area comprises 33607 ha located in Montilla-Moriles D.O., Cordoba (37 • 38 -29 N; 4 • 45 -31 W, 432 m a.s.l.).Montilla is the first production center of Montilla-Moriles wines.This DO produced wines with the grapes of the Pedro Xim énez variety.
The parent material is Triassic marl mainly composed by clays, gypsum and salts.The relief is smooth with slopes ranging from 3 % to 8 %.The most abundant soils are Luvisol (LV) and Cambisol (CM), denominated in the study region "alberos and albarizas", although there are Fluvisol (FL), Regosol (RG) and Vertisol (VR) (FAO, 2006).These soils correspond to the upper limit of the Pliocene period (Andaluciense subperiod), which are characterized by the presence of white marls (argyle-containing limestone) typical of the Guadalquivir basin.These are soft soils owing to the presence of limestone, which combines permeability with high water retention.This latter feature is essential for lands with frequent dry spells.
The Montilla-Moriles DO is characterized by cold winters and warm, dry summers with temperatures ranging from −2 • C to 37.8 • C and an average annual rainfall of 602.7 mm.The moisture regime is dry Mediterranean with continental features due to altitude and location.An unirrigated farm (100 ha) of AC cultivated under conventional tillage (CT) was selected for study in 1965.The soil was a Calcic-Chromic Luvisol (LV cc-cr) (FAO, 2006).In 1966, the study farm (100 ha) was divided into three plots with three different uses (AC, OG and V respectively).The preliminary analyses were realized in 1965 for AC (AC1) and the second analyses were realized in 2011 for AC (AC2), OG and V.In 2011, 22 samples were collected (7 for AC2, 5 for V and 10 for OG) (Figs. 1 and 2).In all cases (AC1, AC2, OG and V) were collected soil entire profiles.Table 1 summarizes the land use class and Table 2 summarize the principal soil properties for the study.Introduction

Conclusions References
Tables Figures

Back Close
Full

Soil sampling and analytical methods
The soil samples were air-dried, ground and sieved through a 2-mm sieve.Soil pH was measured in an aqueous soil extract in deionized water (1:2.5 soil:water) (Guitian and Carballas, 1976).Prior to determining the particle size distribution, the samples were treated with H 2 O 2 (6 %) to remove organic matter.The fraction of particles with a diameter greater than 2 mm was determined by wet sieving.Particles measuring < 2 mm were classified according to USDA standards ( 2004).Soil bulk density was measured by the core method (Blake and hartge, 1986) using a 3.0 cm diameter and 10.0 cm deep core.The distribution of soil particle size was analyzed using the Robinson pipette method (USDA, 2004).SOC were determined by wet oxidation with dichromate according to the Walkley and Black system (1934).TN was determined using the Kjeldahl method (Bremner, 1996).The soil C:N ratio was calculated by dividing the SOC concentration by the TN concentration.The SOC stock (Mg ha −1 ) was calculated for each horizon according to Wang and Dalal (2006) as follows: where d is the thickness of the soil layer (cm), δ 2 mm is the fractional percentage (%) of > 2 mm gravel in the soil, and BD is the bulk density (Mg m −3 ).The TN stock (Mg ha −1 ) was also calculated.The SRs were calculated from SOC, TN and the C:N ratio data following Franzluebbers (2002).The SR is defined as a soil property on the soil surface divided by the same property at a lower depth.In this study, we defined four SRs ([SR1] for Ap/Ap2-Bt, [SR2] for Ap/Bt-B-Ck, [SR3] for Ap/C-Bt-C and [SR4] for Ap/C2-C-Ck).
The statistical analysis was performed using SPSS 13.0 for Windows.The statistical significance of the differences in the variables between land use practices was tested using an Anderson-Darling test at each horizon or a combination of horizons for each soil type.Differences of p < 0.05 were considered statistically significant.Introduction

Conclusions References
Tables Figures

Back Close
Full 3 Results and discussion

Soil properties
The studied soil; a LVcc/cr (FAO, 2006) was characterized by exhibited differences in some physical and chemical parameters with depth.The principal characteristic of these soils is a high clay content in the subsoil (Bt horizon) compared to the topsoil (caused by clay migration) (Table 2).Luvisols are well-developed fertile soils that are suitable for a broad variety of uses typically Mediterranean such as cereals, fruit trees, olives and vineyards (Zdruli et al., 2011).
With respect to the soil thickness, there were no significant differences (p < 0.05) in time (46 yr) for the same land use (AC1 and AC2), however, the LUC from AC1 to V and OG the soil thickness decreased, ranging from 240 cm in AC1 to 182 cm and 176 cm in V and OG respectively (Table 2).This thickness reduction could be caused by the slope steepness, length, topographic curvature and relative position.In this line, McKenzie and Austin (1993) obtained similar results in Australian soils.By contrast, Bakker et al. (2005) in Lesvos-Greece for LUC (AC to V and pastures to OG) between 1956 and 1996 justified this thickness reduction associated with new mechanized equipment (heavy machinery) and water erosion.These causes could be other reasons to justify the thickness reduction in the studied soils.
These soils are characterized by low OM concentrations in depth, especially in V and OG; this can be explained by the soil textures (sandy soils).Gonz ález and Cand ás (2004) found that the formation of OM and mineral aggregates diminishes in the surface horizons of sandy soils, thus favoring high levels of transformed OM, which explain the low OM concentrations at greater depths in the soil studied (Table 2).In addition, Gallardo et al. (2000) explains that the low OM values are explained partly by the semiarid Mediterranean conditions, which are accentuated in Europe's southern soils.Introduction

Conclusions References
Tables Figures

Back Close
Full In all cases (AC1, V and OG), the SOC concentration decreased in depth with the exception of AC2 caused by the high OM concentration in Bt (Table 2).
According to the study of Hernanz et al. (2009) on rainfed crops of Mediterranean semiarid regions; the soil presents a low OC content owing to the high mineralization of the OM and the absence of harvest residues after periods of drought.On the contrary, soils with the coverage of trees show an increase in C and nitrogen (N) (Albretch and Kandji, 2003); we obtained similar results in topsoil, in V and OG.The SOC we found in AC1 was greater (11.1 g kg −1 ) than that estimated by Don et al. (2007), who established 10 g kg −1 for soils with cereal crops in Spain, which must be caused by the accumulation of litter and dead roots in the topsoil.TN and the C:N ratio tended to decrease with depth with the exception of AC2. S á et al. ( 2001) observed an increase in the soil C:N ratio with deep (AC2), which may be attributed to high C:N soluble organic compounds leaching into deeper layers (Diekow et al., 2005).For OG, this decrease may be a result of the increased soil clay content with depth (Table 2).Higher clay content is often associated with more decomposed OM with a lower C:N ratio (Puget and Lal 2005;Yamashita et al., 2006).In the case of AC1 and V, crop residues could favour a higher soil C:N ratio (Puget and Lal, 2005).Additionally, a residue retention can increase the proportion of SOC (Xu et al., 2011) with a lower decomposition degree and higher C:N ratio (Yamashita et al., 2006).Under AC2 the incorporation of residues into the soil can be uniformly distributed with depths up to 20 cm, or more than 20 cm (S á and Lal, 2009;Wright et al., 2007).By contrast, under AC1 and V the input of residues is restricted to the topsoil.Consequently, the soil C:N ratio may be stratified to show a declining trend with depths in the upper soil profile.The C:N ratio in the surface soil was higher than in the lower soil horizons, especially in V (12.23:1) and OG (9.44:1), thus indicating high resolution and separation rates.Lal et al. (1995)  and separation times.Brady and Weil (2008) show that C:N ratios varies between 8:1 and 15:1, with an average of 12:1.

Soil Organic Carbon (SOC) and Total Nitrogen (TN) stocks
The SOC stock for soil groups in Peninsular Spain (FAO soil map of peninsular Spain) is 66.0 Mg ha −1 for LV and soil uses is 50.5 Mg ha −1 , 42.5 Mg ha −1 and 39.9 Mg ha −1 for AC, V and OG respectively (Rodríguez-Murillo, 2001)  and OG (116.7 ± 21 Mg ha −1 ) (Fig. 3).These differences between SOC stocks for soil groups in Peninsular Spain and Andalusia and the study soils are caused by soil thickness (we used soil complete profile -four or five horizons and Rodriguez-Murillo ( 2001) and Mu ñoz-Rojas et al. (2012b) used two or three horizons by profilenot entire profile).
We can be observed that the LUC (AC1 to V and OG) and tillage (AC1 to AC2), reduced the total SOC stock in the long term (46 yr) (Fig. 3).The SOC stored varies within the soil profile, with higher values in Bt horizons for AC1, AC2 and OG, however, in V we found higher SOC in the topsoil.In this line, Novara et al. (2012) for LUC from AC to V obtained similar results and explained that this trend may be due to the mixing of the upper soil layers during soil tillage.SOC stock in the surface horizon in AC1 and AC2 varied from 39.7 Mg ha −1 to 48.9 Mg ha −1 , respectively.Gonz ález and Cand ás (2004) in clayey soils found values near 54 Mg ha −1 in AC.This difference of SOC stock is caused by the texture because our soils were less clayey and sandier (Table 2).According to Burke et al. (1989) and Leifeld et al. (2005), high values of SOC stock in clayey soils are caused by the stabilization mechanisms of the clays in the soil.This effect can be observed in AC1 and OG, which increased the clay content with respect to AC2 and V.By contrast, SOC stored was higher in the subsoil Introduction

Conclusions References
Tables Figures

Back Close
Full (Bt and Bt/C horizons) in AC1 and AC2, which may be due to the translocation of C in the form of dissolved organic C, soil fauna activity, and/or the effects of deeprooting crops (Shrestha et al., 2004).On the contrary, Mu ñoz-Rojas et al. (2012a) for LUC (51-yr period) between 1956 and 2007 in Andalusia (conversion of arable crop to permanent crops) found an increase of SOC for LV (14 %), caused by the limited effect that agricultural management in permanent crops has on SOC sequestration (Smith, 2004).Moreover, Vallejo et al. (2003) indicates that the SOC stock is greater in crop pastures, which is an effect that has also been shown by Nair et al. (2009).Both authors also indicate that the potential for C sequestration in grass systems increases because the roots transfer large amounts of C in the soil slowly and contribute to the increase in the underground C content, which accumulates over time, thus indicating that these systems are more effective in C sequestration than other land uses.
In our study, TN concentrations are elevated in areas where the SOC is high, showing a positive C:N relation (Fig. 3).According to this, clay decreases SOC oxidation and could indicate a positive relationship between clay and nitrogen (Sakin et al., 2010).Some studies (C ôt é et al., 2000) state that the N mineralization decreases when the clay amount increases in the soil.We obtained similar results in LV as TN decreased when the amount of clay increased.McLauchlan (2006) explained that when clay amounts increase in soil, aggregate amounts increase and the potential N mineralization decreases.

The effect of LUC on SOC stock, TN and C:N ratio
A fundamental issue has been to analyze the LUC influence.The change from AC1 to AC2 affected the SOC stock that was reduced 31.2% and the LUC from AC1 to V and OG were reduced 52.7 % and 64.9 %, respectively (Fig. 3).Novara et al. (2012) for LUC from AC to V find an increase of 105 %.Guo and Gifford (2002) reported an increase of 18 % on SOC stock for LUC from AC to plantation.
The loss of SOC stock was caused by tillage (Table 1), AC1 had minimum tillage, with higher biomass production of plantations.As a result, the soil was always covered Introduction

Conclusions References
Tables Figures

Back Close
Full with vegetation, increasing OM stability, which is corroborated by Novara et al. (2012).
The lowest SOC stock was in OG that reduced 64.9 % with respect to AC1.A similar result was obtained by Rodriguez-Murillo (2001) and Padilla et al. (2010), the first for LUC of bushland to OG and the second for LUC of shrubland to OG.This reduction of SOC stock from AC1 to AC2, V and OG, can be explained by a degraded process result of the vegetation losses and an unsustainable soil management resulting in continuous impoverishment in the OM content causing low productivity, which derived in unsuitable chemical properties.Moscatelli et al. (2007) indicated that the SOC loss in cultivated soils could be due to the OM reduced input, as well as to the reduced physical protection of soil from erosion and the increased decomposition rate as a consequence of tillage.The C:N ratio was higher under AC1 than under AC2, V and OG (Fig. 3).This is in line with the results of Blanco-Canqui and Lal (2008) and Lou et al. (2012), which may be explained by the higher contribution of residue input under different tillage.
In OG and V, SOC storage at surface depth (Ap horizon) was high than in the rest of the profile (Fig. 3), West and Post (2002), Puget and Lal (2005), Blanco-Canqui and Lal (2008), argued that SOC stock on surface horizon is greater than in deep due to tillage and in turn increasing the physical protection of native SOC from microbial decomposition.However, AC1 can increase C inputs into surface soil by enhancing crop biomass and in turn residue return.With respect to Bt horizon, this relation was inverse (SOC increased in Bt with respect to Ap horizon).This effect can be explained by soil texture (sandy soils) and tillage, because, native SOC can be reduced on the surface, which may be attributed to soluble organic compounds that can be leaching into deeper layers, increasing the soil aggregates.A similar result was obtained by Diekow et al. (2005).
Franzluebbers and Arshad (1996) and Melero et al. (2009) suggest that minimum tillage can increase SOC under longer experimental duration.We confirmed this because in AC1 increased SOC stock and the LUC and tillage for long-term (40 yr) reduced SOC stock.Introduction

Conclusions References
Tables Figures

Back Close
Full In all cases, the SR of SOC increased in deep with the exception of AC2 (Fig. 4), caused by the low SOC concentration in Ap2/B (transitional horizon between Ap and Bt, caused by the heavy machinery).The SR of SOC for surface to depth [SR1, SR2, SR3 and SR4] increased due to LUC in all situations (AC2, V and OG) (Fig. 4).The LUC improved soil quality because LUC caused alterations in the soil's physical and chemical properties and the soil biotic community (Caravaca et al., 2002).For degraded soils, the SR of SOC is low and occasionally reaches a value of 2.0 (Franzluebbers, 2002).Other studies have shown that SR ranges from 1.1 to 1.9 for conventional tillage (Franzluebbers, 2002;Franzluebbers et al., 2007;Hernanz et al., 2009;S á and Lal, 2009).Higher SR of SOC is a consequence of the accumulation of surface SOC due to straw soil surface coverage and root distribution change.
The SR of TN showed a similar trend in the SR of SOC.The SR of C:N ratio increased in depth, in AC1 and OG, but had no significant differences with respect to soil use.This can be explained by a higher contribution of residue relative to root inputs leading a higher soil C:N ratio (Puget and Lal, 2005).Under AC1, the residue input could have been concentrated on the surface due to straw soil surface coverage, so the soil C:N ratio was stratified.This slight change in C:N ratio suggests the decomposition degree of SOC decreases toward the surface (Lou et al., 2012).This suggests little effect in the LUC and tillage system on the carbon accumulation in the soil.Balesdent and Balabane (1996) do not find any significant differences in SR, in a Geauga farm (Ohio).In AC2, V and OG had ensured the supply of OM from the surface horizons to a deeper horizon, which suggested an accumulation of carbon in the profile under these systems.
The higher soil quality was OG and V, compared to AC1 and A2 was probably due to the presence of a herb layer and the low herbicide applications.The important role of a herb layer, in both protecting soil from the erosion process (Novara et al., 2011) Figures

Back Close
Full and contributing to SOM content, might explain the similarity among the characteristics of the OG and V.

Conclusions
The LUC has a negative impact in the soil, reducing the SOC and TN stocks.The SOC stored varies along the profile, with higher values in the Ap horizon (caused by the mixing of the upper soil layers during soil tillage) and Bt horizons (due to the translocation of C in the form of dissolved organic C, soil fauna activity, and/or the effects of deep-rooting crops).TN concentrations were high in areas where the SOC was high, showing a positive C:N relation.The reduction of SOC by LUC, can be explained by a degraded process (due to vegetation losses and unsustainable soil management, which result in progressive impoverishment in the SOM content, causing low productivity, which derived in unsuitable chemical properties) and by the reduced input of OM in cultivated soils, which reduced physical protection of soil and increased water erosion.However, with respect to the soil quality, 46 yr of LUC had a positive effect in the soil, increasing the SR (in V and OG) of SOC, TN and C:N ratio, caused by the reduction in depth of the SOC and TN.
In general, the LUC reduces the SOC and TN concentrations and by contrast increases the soil quality (SR).The use of entire profiles is necessary in these soils because in temperate climates, large amounts of SOC, may be stored in subsoil horizons.This is essential in LUC because SOC can be transported to deeper soil horizons, contributing to the subsoil C storage.

Conclusions References
Tables Figures

Back Close
Full  Full Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | indicate that C:N ratios are low during resolution Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | C ôt é, L., Brown, S., Par é, D., Fyles, J., and Bauhus, J.: Dynamics of carbon and nitrogen mineralization in relation to stand type, stand age and soil texture in the boreal mixedwood, Soil Biol.Biochem., 32, 1079-1090, 2000.Diekow, J., Mielniczuk, J., Knicker, H., Bayer, C., Dick, D. P., and K ögel-Knabner, I.: Soil C and N stocks as affected by cropping systems and nitrogen fertilisation in a southern Brazil Acrisol managed under no-tillage for 17 years, Soil Till.Res., 81, 87-95, 2005.Don, A., Schumacher, J., Scherer-Lorenzen, M., Scholten, T., and Schulze, E.: Spatial and vertical variation of soil carbon at two grassland sites.Implications for measuring soil carbon stocks, Geoderma, 141, 272-282, 2007Discussion Paper | Discussion Paper | Discussion Paper | S á, J. C. M., Cerri, C. C., Dick, W. A., Lal, R., Venske-Filho, S. P., Piccolo, M. C., and Feigl, B. E.: Organic matter dynamics and carbon sequestration rates for a tillage chronosequence in a Brazilian Oxisol, Soil Sci.Soc.Am.J., 65, 1486-1499, 2001.S á, J. C. M. and Lal, R.: Stratification ratio of soil organic matter pools as an indicator of carbon sequestration in a tillage chronosequence on a Brazilian Oxisol, Soil Till.Res., 103, 46-56Discussion Paper | Discussion Paper | Discussion Paper |