Soil and crop management practices and the water regulation functions of soils: a qualitative synthesis of meta-analyses relevant to European agriculture

. Adopting soil and crop management practices that conserve or enhance soil structure is critical for supporting the sustainable adaptation of agriculture to climate change, as it should help maintain agricultural production in the face of increasing drought or water excess without impairing environmental quality. In this paper, we evaluate the evidence for this assertion by synthesizing the results of 34 published meta-analyses of the effects of such practices on soil physical and hydraulic properties relevant for 5 climate change adaptation in European agriculture. We also review an additional 127 meta-analyses that investigated synergies and trade-offs or help to explain the effects of soil and crop management in terms of the underlying processes and mechanisms. Finally, we identify how responses to alternative soil-crop management systems vary under contrasting agro-environmental conditions across Europe. This information may help practitioners and policymakers to draw context-specific conclusions concerning the efficacy of management practices as climate adaptation tools. 10 Our synthesis demonstrates that organic soil amendments and the adoption of practices that maintain “continuous living cover” result in significant benefits for the water regulation function of soils, mostly arising from the additional carbon inputs to soil and the stimulation of biological processes. These effects are clearly related to improved soil aggregation and enhanced bio-porosity, both of which reduce surface runoff and increase infiltration. One potentially negative consequence of these systems is a reduction in soil water storage and


Introduction
As a consequence of on-going climate change, the occurrence of extreme weather events (i.e. heatwaves, summer droughts, waterlogging and flooding) such as those experienced during the recent summers of 2018, 2021 and 2022 will almost certainly increase in all parts of Europe IPCC (2021); AgriAdapt (2017). Climate change impacts on agriculture are projected to result in 30 an average 1% loss of gross domestic product by 2050, but with large differences among regions and farming systems Jacobs et al. (2019). An urgent task is therefore to develop guidance on soil and crop management practices that would help farmers in all regions of Europe adapt to these extreme weather situations.
The ecosystem services a soil can deliver depend profoundly on its structure, which we define here as the spatial arrangement of the soil pore space. Mediated by various biological (e.g., faunal and microbial activity) and physical processes (e.g., traffic 35 compaction, wet-dry and freeze-thaw cycles), soil structure is constantly evolving at time scales ranging from seconds to centuries, driven by weather patterns as well as changes in climate and land management practices ( Figure 1). In turn, soil structure strongly affects all life in soil as well as the balance between infiltration and surface runoff, as well as drainage and soil water retention and therefore the supply of water and nutrients to crops. Agricultural practices can affect soil structure directly (e.g., compaction due to use of heavy machinery) or indirectly (e.g. improved soil structure through increased bioturbation 40 by earthworms after addition of organic matter to the soil). Practices commonly adopted in "conservation agriculture " Palm et al. (2014) are thought to enhance soil structure and should therefore help to maintain agricultural production in the face of severe droughts or heavy rain. Conservation agriculture to improve soil structure rests on three fundamental principles Palm et al. (2014): i.) minimizing mechanical soil disturbance, ii.) maintaining soil cover by plants as much as possible and for as long as possible (i.e. aspects of both spatial and temporal coverage), and iii.) diversifying cropping. Other more recently 45 coined and partially related terms are "regenerative agriculture", which acknowledges past failures to preserve soil health Schreefel et al. (2020) and "climate-smart agriculture", which is defined by FAO (2010) as ". . . agriculture that sustainably increases productivity, enhances resilience, reduces greenhouse gases, and enhances achievement of national food security and development goals".
The effects of soil and crop management practices on soil properties, soil hydrological and biological functioning and crop 50 performance have been studied in many long-term field trials throughout the world. In addition to narrative reviews (e.g. Palm et al. (2014)), many quantitative meta-analyses synthesizing the findings of individual experiments have also been published. This is especially the case in the last few years Beillouin et al. (2019b, a), probably because the number of field experiments that have been running for a sufficient length of time has only recently reached the critical mass required to enable these kinds of quantitative analyses. Indeed, the increase in the number of meta-analyses published on topics related to conservation agriculture 55 has been so dramatic that four over-arching syntheses of these meta-analyses have also recently been published. Bolinder et al. (2020) evaluated the effects of organic amendments and cover crops on soil organic matter (SOM) storage, while Schmidt et al. (2021) focused on the effects of biochar on crop performance. Beillouin et al. (2019b) and Tamburini et al. (2020a) carried out even more ambitious and comprehensive reviews of meta-analyses of the effects of conservation agriculture and crop diversification strategies on a wide range of ecosystem services. Tamburini et al. (2020a) concluded that diversification 60 practices most often resulted in a 'win-win' situation for ecosystem services including crop yields, but that the large variability in responses and the occurrence of trade-offs highlighted the need to analyze the context-dependency of outcomes, something which was only possible to do to a limited extent with their broad-brush treatment. These previous syntheses of meta-analyses on the benefits of conservation agriculture have placed very little emphasis Tamburini et al. (2020b) or none at all Bolinder et al. (2020); Beillouin et al. (2019b, a) on soil hydrological functioning even though this is key for climate change adaptation. In 65 their synthesis, Tamburini et al. (2020a) included 17 meta-analyses (involving 31 effects-size comparisons) relevant to water regulation, but most of these concerned water quality issues rather than hydrological functioning per se. Beillouin et al. (2019b) concluded that . . . "our review reveals that a significant knowledge gap remains, in particular regarding water use".
In this study, we focus on the implications of agricultural management practices for soil hydrological functioning for climate change adaptation under European agro-environmental conditions. We do this by identifying and synthesizing existing 70 meta-analyses of the response of soil physical/hydraulic properties and hydrological processes relevant for climate change adaptation to soil and crop management practices. In those cases where the information is available, we summarize knowledge of context-specific effects of relevance for the range of agro-environmental conditions found in Europe, and as far as possible, explain these variations in terms of individual driving processes and mechanisms. This kind of information may explain local praxis in agricultural management (i.e. farmer behavior) and will enable practitioners and policymakers to draw context-specific 75 conclusions concerning the efficacy of management practices as climate adaptation tools. This study highlights where consensus has been established on practices improving the water regulation function of soil that are meaningful for climate change adaptation. We also identify remaining knowledge gaps and key avenues for future research. The text string shown in Figure 2 was used to search the published literature using Web of Knowledge in May 2021. This search returned 663 results. All search results were manually assessed for their relevance to the objectives of our study. Meta-analyses that only included studies carried out outside Europe were not retained. Our search identified 34 relevant meta-analyses focusing on the effects of soil and crop management on soil physical properties and hydrological processes using effects ratios (Appendix B). Figure 3 shows the number of primary studies per publication year included in the 34 meta-analyses. A peak is clearly visible 85 4 Figure 2. Search string used to identify relevant meta-analyses in 2014, which is explained by the fact that all the selected meta-analyses were published after 2015. Our search string was also designed to identify meta-analyses of management effects on soil organic matter and biological variables (e.g. microbial biomass), since these help to explain the observed effects on physical/hydraulic properties and hydrological processes, as well as other studies that analyzed target variables representing potential "trade-offs" or synergies. Among these, we focused primarily on the impacts of management practices on crop yields, greenhouse gas emissions and water quality. An additional 127 published 90 meta-analyses of this kind were identified by our literature search. These studies are listed in the supplementary file ("Supporting studies.xlsx").
The target variables (e.g. soil physical and hydraulic properties) and drivers (i.e. soil and crop management practices) included in the 34 meta-analyses were then classified into a limited number of groups. The target variables were grouped into five classes: pore space properties (e.g. porosity, bulk density), hydraulic properties (e.g. saturated hydraulic conductivity, field capacity), 95 mechanical properties (e.g. soil aggregate stability, penetration resistance), water flows (e.g. infiltration, surface runoff, drainage) and plant properties (e.g. root length density, water use efficiency). Likewise, the management practices were also grouped into five classes: soil amendments (e.g. manure, biochar, organic farming systems), cropping practices and systems (e.g. cover crops, crop rotations), tillage systems (e.g. no-till), grazing management and irrigation.

100
We performed a quality assessment of the selected 34 meta-analyses using 15 of the criteria proposed by Beillouin et al. (2019a). Figure 4 presents a summary of the quality of the selected meta-analyses according to these criteria. Nearly half of the meta-analyses included datasets in the paper, while only ca. 44% investigated publication bias Philibert et al. (2012). The authors of these studies used simple statistical techniques such as frequency distributions of effects sizes or "funnel plots" of sample sizes against effect sizes to investigate whether experiments with non-significant effects are under-represented in the

Redundancy analysis
We performed a redundancy analysis to identify the proportion of common primary or source studies among the meta-analyses following the methodology of Beillouin et al. (2019a). For each of the 34 selected meta-analyses, the references to the studies used were extracted from the supplementary materials. Each reference contained at least the name of the first author, the year of publication, the title, the journal and -if available -the DOI. Of the 3142 unique primary studies, 437 had no DOI. Old 115 publications or publications not written in English were usually found to have no DOI. In some cases, the title and DOI were not available so we had to manually check these references based on contextual information supplied in the supplementary material. In most cases, however, the title was provided in the meta-analysis and the DOI could be extracted automatically from the Cross-Ref database. We then manually checked if the title of the paper matched the one found on Cross-ref, to confirm the DOI assignment. The results of the redundancy analysis are presented in the Appendix A as well as in the notebook 120 6 Figure 4. Proportion of the quality criteria defined by Beillouin et al. (2019a) that are met by the selected meta-analyses in this study at https://github.com/climasoma/review-of-meta-analyses/blob/main/notebooks/redundancy.ipynb. The main outcome of this analysis is that redundancy is only problematic for a few meta-analyses on biochar that were published almost simultaneously ; .

Qualitative analysis of effect sizes
In total, the 34 meta-analyses reported 104 effects ratios comparing the impacts of a management practice to a control treatment 125 for a particular response variable. The overall effects of these treatments on the target variables (either positive, negative or neutral i.e. non-significant) were read from tables and figures in each of the 34 meta-analyses and analyzed in a qualitative way. This is because we do not have access to the individual effects sizes in all the primary studies included in the meta-analyses (see Section 2.2 "Quality assessment"). We considered the effect as "positive" if the average log response ratio and the entire 95% confidence interval reported in the meta-analysis was larger than zero (equivalent to a response ratio of 1 prior to taking 130 logarithms). If part of the 95% confidence interval for the log response ratio overlapped zero, the effect was considered "neutral".
If the entire confidence interval was smaller than zero, the overall effect was considered as "negative". The directions of the effect sizes are therefore purely statistical and have no connotation of value. We report the effects in a statistical sense because in some instances it would not be clear whether effects would be beneficial or detrimental. It should also be noted that positive or negative overall effects derived collected from a given meta-analysis do not imply that all the individual effects in the primary 135 studies included in this meta-analysis necessarily pointed in the same direction. For all overall effects retrieved, we also noted the number of individual effects from the primary studies used to compute the overall effect reported in the meta-analysis. 7 3 Results and discussion Figure 5 summarizes the statistical relationships found between the drivers and target variables in the selected meta-analyses. It shows that the effects of cropping systems, tillage, organic amendments and, to a lesser extent, irrigation management have been 140 studied extensively. These topics are discussed in the following sections. It is equally interesting to consider the empty zones in Figure 5, which represent topics for which existing experimental data has not yet been summarized or which have been the focus of only a few studies in the past. We discuss these knowledge gaps in Section 3.5. Finally, we use the outcome of our analysis to outline some key avenues for future research on the extent to which management practices can reinforce the water regulation function of soils.

Cropping systems and practices
Broadly speaking, published meta-analyses that have investigated the effects of cropping systems and practices ( Figure 5) fall into two categories: i.) studies analyzing the effects of maintaining a more continuous soil surface cover, either in a temporal (e.g. cover crops in arable rotations) or in a spatial sense (e.g. inter-row cover in widely-spaced row crops such as vineyards and orchards), and ii.) studies comparing farming systems (e.g. continuous arable contrasted with either perennial crops or rotations 150 or mixed farming systems with livestock). In the following, we combine these two aspects, referring to both of them as cropping systems that as far as possible maintain a "continuous living cover" Basche and deLonge (2017). increased proportion of perennial crops in the rotation and the presence of ground cover between the rows of perennial crops (e.g. in vineyards) increase soil infiltration and reduce surface runoff ; Liu et al. (2021); Basche and deLonge (2019). These positive effects seem broadly similar regardless of climate ; Liu et al. (2021). As noted above, "continuous living cover" increases soil organic matter contents and both long-term field experiments and meta-analyses suggest that soil organic matter generally tends to increase the plant available water capacity. However, although the magnitude of this 175 effect is still a matter of debate Lal (2020), in most cases it seems relatively small compared with the crop water demand Minasny and McBratney (2018b); Libohova et al. (2018); Minasny and McBratney (2018a). One potential negative effect of cropping systems employing "continuous living cover" is that increased transpiration may reduce soil water contents Shackelford et al. of other field operations and the lack of mechanical loosening by cultivation Hamza and Anderson (2005). Peixoto et al. (2020) showed that these negative effects can be alleviated with occasional tillage. Soil tillage indirectly affects soil structure through effects on soil macro-fauna. In addition to the direct impacts of tillage implements on mortality, disruption of the soil also exposes soil macro-fauna to increased risks of desiccation and predation.
Consequently, meta-analyses show that total earthworm biomass and abundance increase as tillage intensity is reduced Spurgeon  (2017), with a negative relationship between tillage depth and earthworm abundance Briones and Schmidt (2017). Deep burrowing and surface-feeding (anecic) earthworm species are particularly favored by NT systems, as their permanent burrows are no longer destroyed by ploughing and they have a better access to food resources.
Thus, a lack of disturbance of the soil by tillage has also been shown to increase the diversity of earthworm populations Spurgeon et al. (2013); Briones and Schmidt (2017); Chan (2001) and soil fauna in general Graaff et al. (2019).  Li et al. (2020a). This suggests that the effects of the enhanced bioporosity in NT systems created by soil fauna, and especially anecic earthworms, on saturated and near-saturated hydraulic conductivity Lee and Foster (1991) generally outweigh the negative effects of increased bulk density. Thus, Spurgeon et al. (2013) showed that increased earthworm abundances and diversity found under NT systems were positively correlated with infiltration rates. Comparing ecological groups, they found that the density of anecic earthworms was positively associated with increased infiltration rates, whereas no effect was apparent 235 for endogeic earthworms. In principle, better-developed soil macropore systems and improvements in aggregate stability and infiltration capacity should promote a more favorable crop water balance, with reductions in surface runoff. Figure 5 shows that the effect on runoff is one of the most studied hydrological processes related to tillage. The meta-analysis performed by  found that RT and NT systems decreased surface runoff. However, these results do not appear to be conclusive as two later meta-analyses ;  failed to detect significant effects of conservation tillage 240 practices on surface runoff. However,  found that contour tillage and deep tillage both reduced surface runoff.

Synergies and trade-offs
Adoption of no-till and reduced tillage systems involve several trade-offs, particularly concerning water quality, GHG emissions and crop yields. NT systems tend to give smaller yields for many crops compared with conventional tillage Mangalassery et al. showed that occasional tillage increased crop yields compared with NT in dry regions and in soils with limited water retention capacity and availability, presumably by alleviating soil compaction and improving rooting.
With respect to water quality, Daryanto et al. (2017a) found an overall 40% reduction in phosphorus loads in surface runoff for NT systems in comparison with CT. This was attributed to significant decreases in losses of particulate phosphorus, as concentrations of dissolved P actually increased in runoff under NT. For pesticides, Elias et al. (2018) found no significant differ-255 ences in concentrations in surface runoff for 14 of the 18 compounds included in their meta-analysis. Pesticide concentrations were actually larger under NT for the remaining 4 compounds. For loads, no significant difference was detected between CT and NT systems for 15 of the 18 pesticide compounds. For the three remaining pesticides, losses in surface runoff were larger under NT for metribuzin and dicamba and smaller for alachlor. As also noted by Elias et al. (2018) , these results seem quite surprising given the documented effects of conservation tillage on soil structure and hydraulic properties in the uppermost soil 260 layers discussed earlier, which should increase soil infiltration capacity and reduce surface runoff. For nitrate losses in surface runoff in conventional and no-till systems, Daryanto et al. (2017b) showed that a change to NT resulted in an increase in nitrate concentrations in surface runoff, but similar loads, implying that surface runoff was, as expected, less prevalent under NT. Daryanto et al. (2017b) also performed a meta-analysis on nitrate leaching. They found larger leachate losses of nitrate under NT systems than CT, whereas the concentrations in leachate were similar under both tillage systems, indicating that the effect 265 of NT on nitrate leaching was largely determined by increases in water percolation. We did not find any meta-analyses on the effects of tillage systems on pesticide leaching in our literature search. Leaching is the outcome of several interacting processes involving many complex and poorly understood processes Alletto et al. (2010). In practice, with no mechanical disturbance, larger quantities of pesticides are often used to control weeds and diseases in NT systems. However, pesticide leaching will also be highly sensitive to changes induced by tillage in soil structure, microbial biomass and activity and SOC, since these will affect 270 water flow velocities, degradation rates and the strength of adsorption in soil. Several studies suggest that the better-preserved macropore networks established under RT and NT systems may enhance leaching by preferential flow Jarvis (2007) Alletto et al. (2010). Although it is difficult to draw firm conclusions about the effects of conservation tillage practices on pesticide leaching without the help of quantitative meta-analyses, we may tentatively conclude that the greater risk of macropore flow under RT and NT systems appears to outweigh any beneficial impacts of increases in SOC and microbial 275 activity on pesticide adsorption and degradation.
Significant trade-offs have also been reported with respect to greenhouse gases. In an early meta-analysis, van Kessel et al. (2013) found no overall impact of reduced tillage or no-till on N 2 O emissions, with observed increases in humid climates compensated by reductions in emissions in drier climates, although neither trend was significant. However, in a later metaanalysis, Mei et al. (2018) reported a significant overall increase of 18% in N 2 O emissions under conservation tillage, with the 280 largest effects in warmer and wetter climates and in finer-textured soils. Shakoor et al. (2021) found significant increases of emissions of CO 2 , N 2 O and CH 4 of 7, 12 and 21% respectively under NT compared with CT. From the perspective of climate change mitigation, Guenet et al. (2020) concluded that increased greenhouse gas emissions under NT outweighed any minor gains in soil C stocks.

Biochar
Biochar is charcoal made for the purpose of soil amendment. It is a type of black carbon, resulting from incomplete combustion of organic matter through pyrolysis. Apart from its potential for long-term soil carbon sequestration, it can also have beneficial effects on nutrient availability and soil physical properties Joseph et al. (2021). The quantitative analysis of the effects of biochar on physical and hydraulic properties shown in Figure 5 is based on effects ratios presented in five meta-analyses Rabbi et al.  2021) only presented data for various sub-categories (e.g. for different types of biochar) and not for the overall effects of biochar addition. Taken altogether, these seven studies present results of analyses for different soil types, textural classes and experimental conditions (i.e. field or laboratory/greenhouse study) as well as for biochars of different properties and applied at 295 different rates. These analyses show that biochar has several positive effects on soil hydraulic properties, but that these effects are dependent on all of the above-mentioned variables.
Decreases in bulk density and increases in porosity are generally reported after biochar addition . The density of biochar is low and the porosity is often high compared to soil, which may explain the observed effects. However, if biochar mainly fills existing pores, porosity will decrease and bulk density increase. Biochar will also 300 influence these variables indirectly through its effects on aggregation Pituello et al. (2018). Two meta-analyses reported measures of aggregate stability.  included only studies that reported mean weight diameter (MWD) using wet sieving while Islam et al. (2021) included studies that reported soil aggregate stability as a percentage of water-stable aggregates (WSA), as well as MWD or gravimetric mean diameter (GMD) using either wet sieving or dry sieving. Both studies showed that 13 aggregate stability increased with biochar addition and that these effects increased with the time between biochar application 305 and measurements Islam et al. (2021). Figure 5 suggests that biochar addition generally increases the plant available water content (θ paw ). These meta-analyses show that although the water contents at field capacity (θ fc ; pressure potentials in the range between -0.033 and -0.01 MPa) and wilting point (θ pwp ) both tend to increase following biochar amendment, the effects on θ fc appear to be larger ( Figure 5 ). Pore sizes in biochars range over at least five orders of magnitude, from the sub-nanometer scale to pore diameters of the order of 310 tens of micrometers originating from partially preserved cellular structures Brewer et al. (2014). However, a large fraction of the pore volume in biochar consists of pores in the nanometer size range Downie et al. (2009). These pores will retain water at very low pressure potentials and therefore increase the wilting point water content θ pwp upon biochar addition. It has been suggested that increases in θ paw may be due to the filling of existing soil macropores with biochar, which would shift the pore size distribution from large pores that drain quickly to pores that can retain water at field capacity Liu et al. (2017). Biochar 315 itself contains pores in the relevant size range (0.2-100 µm in diameter) to contribute to θ paw . Thus, inter-particle pores in biochar will also contribute to θ paw depending on the size distribution and shapes of the biochar particles and their effects on soil aggregation Burgeon et al. (2021). Since θ fc is the sum of θ pwp and θ paw , the same processes are the likely causes of the observed increases in θ fc .
The effects of biochar on water retention were in most cases larger for coarse-textured soils. Biochar with large microporosity 320 can fill the larger inter-particle soil pores present in sandy soils so that the pore size distribution shifts towards the smaller pores that can retain water at the pressure potentials corresponding to field capacity ; . Moreover, fine-textured soils retain more water at θ fc so that the relative changes induced by biochar may be smaller .
All the meta-analyses included data on the effects of biochar production parameters (e.g. feedstock, pyrolysis temperature) and 325 the chemical and physical properties of biochar. Generally, the influence of these parameters on the effects of biochar addition were minor with respect to soil water retention. Due to lack of data, the influence of the time between biochar application and measurements on the effects on water retention was not included. It is, however, clear from studies on century-old charcoal kiln sites that the properties of biochar and associated soil evolve over time Cheng et al. (2008); Hardy et al. (2016).
One meta-analysis reported an increase in saturated hydraulic conductivity following biochar addition Omondi et al. (2016), 330 while two others reported negative effects ( Figure 5). Saturated hydraulic conductivity is a function of pore network properties, including connectivity of the macropores and the presence of pore bottlenecks Koestel et al. (2018). A few studies have quantified the effects of biochar addition on the connectivity of macropore networks using X-ray tomography Yan et al. (2021); Yu and Lu (2019). These studies indicate that the connected macroporosity and the diameter of pore throats decrease in mediumto coarse-textured soils amended with biochar. However, the influence of soil texture on the effects of biochar on saturated 335 hydraulic conductivity reported in the meta-analyses is not consistent.
The effects of biochar on the studied variables were in many cases larger for higher application rates. Often, laboratory studies used much larger application rates (>50 t ha -1 ) than field studies and also reported larger effects. Additionally, as pointed out by , mixing of biochar after field applications is challenging and may be another reason why effects were sometimes small or insignificant for field experiments. The majority of the studies included in the meta-analyses were short-term 340 experiments (i.e. duration <1 year). Future work should therefore focus on longer-term effects of biochar applications under realistic field conditions. This requires either long-term field experimentation, which is expensive, or the study of historical biochar sites. The meta-analyses show large variations in effects for all the included variables, which suggests that future work should also be directed towards finding biochars with specific properties (e.g. surface area, particle size) designed to improve soil physical properties under specific soil and climate conditions while maintaining or improving nutrient availability. 345 3.3.2 Other organic amendments, residue retention and mulching Figure 5 shows that only a few meta-analyses have focused specifically on the effects of organic soil amendments or residue retention and mulching on soil properties relevant for water regulation functions. Instead, these practices are often included in meta-analyses on conservation agriculture or tillage systems. In these studies, the effects of the treatments are combined.
Furthermore, the influence of contrasting soils or climates has not been assessed. and hydraulic properties. They found that aggregate stability increased with organic amendments and that this effect was largest for compost. However, this beneficial effect decreased with time.  found that application of soil amendments reduced both surface runoff and soil erosion. Gravuer et al. (2019) analysed effects of organic amendments (manure, biosolids and compost) applied to arid, semi-arid and Mediterranean rangelands. They found increased water contents at field capacity 355 and reduced surface runoff. Additional benefits were increased soil organic carbon contents and above-ground net primary productivity, while trade-offs were increased CO 2 emissions, increased soil lead concentrations and increased losses of N and P in surface runoff.
Mulching means to add (or retain) material on the soil surface without incorporation Kader et al. (2017). In this review, we focus on organic mulches, but synthetic materials are also used. The most extreme example of mulching with artificial materials 360 is plastic mulching, which has been shown to increase crop water efficiency under drought Yu et al. (2021). The use of organic amendments may have several beneficial effects on soil quality and the environment and is therefore one important practice in conservation agriculture. Mulching is typically carried out to limit soil evaporation, reduce soil runoff and erosion but it also affects, among other things, nutrient cycling, weed infestations and soil carbon storage Ranaivoson et al. (2017). Mulching was included as one driver in four meta-analyses that studied effects on soil hydraulic functions. These meta-analyses showed 365 positive effects on the rather limited number of hydraulic properties included ( Figure 5). Three meta-analyses analyzed the effects of mulching on surface runoff, one for agricultural land , one for annual crops Ranaivoson et al. (2017) and one focusing only on tree crops Liu et al. (2021). They all showed reduced surface runoff. The study for annual crops also showed reduced soil evaporation and increased infiltration. These effects are already well-established in the scientific literature and in line with the intentions of mulching Kader et al. (2017). The meta-analysis by  focused on effects of 370 different tillage practices which also includes the comparison between residue retention and removal in no-till systems. Residue retention led to a decrease in bulk density, an increase in total porosity and an increase in plant available water whereas it did not have significant effects on saturated hydraulic conductivity. They attributed this to increased accumulation of organic material on the soil surface, which leads to increased biological activity and soil aggregation.
Overall, organic amendments have potentially beneficial effects for several soil properties relevant for water regulation.

375
Although this is not a new observation, techniques such as mulching or biochar application are still rather little applied. Residue retention is more common in the EU, but worldwide residues are often still burnt on the field or collected for other uses. The limited availability of the organic material at the right place at the right time and for an acceptable price is most probably one of the major bottlenecks for a widespread application of these amendments, especially in the case of biochar. Future research in this field should therefore urgently tackle the socio-economic challenges related to the availability of organic amendments that 380 prevent their widespread use.  . The objective of this approach to irrigation scheduling is to reduce water use without significantly impacting 385 yields by limiting the supply of water during periods of the growing season when it is less critical for crop growth. One of these meta-analyses  also synthesized the results of studies investigating the effects of partial root zone irrigation on water use efficiency and crop yields. This method also has the objective of saving water without impacting yields, but in this case by alternately suppling water to only one part of the root zone at each irrigation. These meta-analyses show that although these irrigation scheduling methods either have mostly neutral or sometimes positive effects on crop water use 390 efficiency ( Figure 5), crop yields are significantly smaller compared to full irrigation for almost all crops and soil types. This implies that crop yields may in some cases be reduced less than water consumption, although these water savings may not compensate farmers for their yield losses. Another way to conserve high quality fresh water resources is to make use of brackish or saline water for irrigation. In their meta-analysis,  showed how the decreases in water productivity, irrigation use efficiency and crop yields as a result of the use of salty irrigation water ( Figure 5) depends on crop type, irrigation 395 methods, climates and soil type. Figure 5 shows that only one meta-analysis has studied the effects of management on crop root system characteristics. This is most probably because root system characterization is tedious, time-consuming and too invasive for long-term field trials.

Knowledge gaps
Nevertheless, the soil-root interface is a crucial environment mediating the flow of water in the soil-plant-atmosphere system. It 400 can be expected that many soil management practices influence root penetration and rooting depths, thereby strongly influencing potential rates of water uptake by plants during droughts. At present, we can only make inferences about the effects of soil management on crop transpiration, either from other terms in the soil water balance or from the use of penetration resistance as a proxy.
We also showed that although several recent meta-analyses have investigated the impacts of different irrigation scheduling 405 strategies on water use efficiency and crop yields, none has so far summarized the effects of irrigation on soil physical properties.
Nevertheless, the type of irrigation technique (e.g., surface, sprinkler or drip irrigation) and the quality of water used are known to strongly affect soil structure and hydraulic properties Sun et al. (2018); Leuther et al. (2019); Drewry et al. (2020), which should impact the water regulation functions of soil. A quantitative summary of existing experimental information would provide critical support to policies and practices for effective adaptation of farming systems to future climates with more frequent and 410 severe summer droughts.
We would also like to highlight some additional knowledge gaps that are not revealed in Figure 5. Firstly, with some exceptions (i.e. tillage practices and residue management), most long-term field experiments only have simple designs that neglect some potentially interesting combinations of treatments (e.g. the interactions between soil and crop management and irrigation systems). Secondly, some key target variables are rarely measured and so have not yet been the subject of meta-analysis.

415
For example, most long-term field trials on the effects of soil and crop management practices on hydrological functioning have measured proxy variables for soil structure, such as infiltration or soil hydraulic properties (water retention, hydraulic conductivity at and near saturation). No meta-analyses have been performed yet for metrics quantifying various aspects of soil structure per se Rabot et al. (2018) even though the application of X-ray imaging techniques to quantify soil structure is becoming increasingly common. As a result, the number of X-ray studies published is rapidly increasing, so it should soon be 420 possible to carry out such an analysis.

Conclusions and outlook
A large number of meta-analyses have been published in recent years on the impacts of soil and crop management practices on soil properties and processes and the various ecosystem services and functions delivered by soil. In this report, we have synthesized these analyses with respect to the water regulation functions that are relevant for climate change adaptation in Europe.

425
Across Europe, climatic extremes (i.e. droughts and intense rains) will become more frequent and more severe. Specifically, effective adaptation to climate change requires soils with a well-developed and stable structure with a large infiltration capacity and an ability to sustain water supply to plants during extended dry periods. This synthesis has revealed a considerable degree of consensus concerning the effects of soil and crop management practices on key soil properties relevant for these hydrological functions.

430
Meta-analyses have demonstrated that the use of organic amendments and the adoption of cropping systems and practices that maintain, as far as possible, "continuous living cover" both result in significant beneficial effects for the water regulation function of soils, arising from the additional carbon inputs to soil and the stimulation of biological processes. These effects are clearly related to improvements in soil structure, both in terms of stable aggregation at the micro-scale and enhanced bio-porosity, both of which reduce surface runoff and increase infiltration. One potentially negative consequence of management practices that 435 maintain "continuous living cover" is a reduction in soil water storage and groundwater recharge. This may be problematic in dry climates, where there is some evidence to suggest that yields of the main crop may be affected. With respect to environmental quality, no other significant trade-offs are known, while some important synergies have been identified, in particular reductions in nitrate leaching to groundwater and greenhouse gas emissions.
The amelioration of soil structure that occurs under reduced (RT) and no-till (NT) practices may improve infiltration capacity 440 and reduce surface runoff, despite the increases in bulk density that are commonly reported, although the evidence for this is inconclusive. Furthermore, some significant trade-offs with RT and NT systems have also been identified. For example, yield penalties incurred under NT and increased weed pressure and/or increased herbicide use and thus leaching risks, especially in wetter and colder climates, constitute a barrier to adoption by farmers. Greenhouse gas emissions are also generally larger under NT, while leaching losses to groundwater of both nitrate and pesticides may also increase. Although we might expect losses of 445 agro-chemicals in surface runoff to generally decrease under RT and NT, thereby compensating for greater leaching losses, this does not always appear to be the case. Reduced tillage intensity in the temporal sense (i.e. "occasional" tillage) may help to ameliorate some of the negative effects of no-till systems, whilst retaining some of the advantages.
Our extensive synthesis of the existing literature has also identified several important knowledge gaps, particularly related to the effects of management practices on soil structure, root growth and transpiration and on combinations of practices. Thus, 450 conclusions related to the impacts of management on the crop water supply are necessarily based on inferences derived from proxy variables such as available water capacity and infiltration capacity.
To address these limitations, we recommend that future research should focus on the following: 1. monitoring transpiration (e.g. by sap flow) and crop root development in existing field trials and the development of techniques to do this in a minimally invasive way for the entire soil root zone; 455 2. monitoring of soil structure and hydraulic properties in field trials over the entire soil profile; 3. application of soil-crop models making use of measured hydraulic properties and climate model projections to evaluate and predict the impacts of alternative soil/crop management practices on water balance and crop yields under climate change; 4. introduction of irrigation and drought treatments at existing long-term field trials to investigate the consequences for water 460 regulation functions under climate change.

Author contribution
Conceptualization: all co-authors Data collation and analysis: GuB, GiB, NJ Writing of the original draft: GuB, NJ, ML, KM, SG 465 Writing, review and editing: all co-authors

Competing interests
No competing interests are present.

Data availability
Data is available at https://github.com/climasoma/review-of-meta-analyses Acknowledgements This work was funded by the EU EJP SOIL project CLIMASOMA ("Climate change adaptation through soil and crop management: synthesis and ways forward") with H2020 Grant agreement number: 862695.

475
Appendix A: Redundancy analysis Note that for this analysis, the studies of  and Li et al. (2020b) were considered as one, as they both rely on the same database, but analyze different variables. We first identified the studies shared between multiple meta-analyses and computed the percentage of shared studies per meta-analysis. Figure A1 shows the percentage of shared studies (number of shared studies divided by number of studies in the meta-analysis in the row times 100). A2 shows for each meta-analysis the 480 number of source studies that it shares with at least one other meta-analysis. Some meta-analyses share nearly 100% of their studies with another meta-analysis (e.g. . In addition to the extent of redundancy, A2 also shows the number of primary studies included in each meta-analysis.   20 Figure A3. Redundancy among the selected meta-analyses (horizontal axis). Dots represent the percentage of shared primary studies between two meta-analyses. When this percentage is above 25%, the dots are shown in red, and the name of the meta-analysis is displayed. For instance, Li et al. (2020) shares more than 25% of its primary studies with the meta-analysis of Mondal et al (2020). The number on the horizontal axis denotes the number of other meta-analyses that share primary studies with the meta-analysis named horizontally. Note that several meta-analysis do not share any studies with others. Meta-analysis are sorted according to the amount of shared primary studies they have (same order as A2).