Search

Soil moisture (SM) and temperature (ST) are critical factors in forest eco-hydrological research. In this study, we investigated the inter- and intra-annual changes in SM and ST profiles in a mixed Mediterranean maquis forest stand together with soil and meteorological parameters. Hourly data from three field measurements points at four depths (−5, −20, −40 and −70 cm) for 6 years were interpolated using the kriging method to produce annual SM and ST profiles. The results indicate that air temperature highly affects the upper 5 cm of the mineral soil. In general, it increases with depth in winter at an average rate of 0.036 °C/cm and decreases in summer (0.035 °C/cm), presenting higher values compared to air temperature from April to August and lower ones during the rest of the period. Precipitation is the main factor driving SM variations up to a superficial soil depth of 40 cm. The upper soil layer (0–40 cm) infiltrates water faster and presents high SM variability, especially in monthly and seasonal (year to year) time steps. The maquis forest stands are likely to be strongly affected by climate change, therefore the results of this study could be useful in hydrological and climate change studies focused on maquis vegetation water management. [ABSTRACT FROM AUTHOR]

Leaf respiration drives plant energy use and greatly influences global carbon balance. It is inhibited by light. The climate-driven effects of changes in nitrogen deposition and soil moisture on leaf respiration in light (RL) are not well understood. To better understand the response of RL and its inhibition by light, we experimentally determined RL and light inhibition degree, an index of the inhibitory effect of light on leaf respiration with respect to dark leaf respiration, for potted specimens of an evergreen conifer species (Pinus koraiensis) and a deciduous broadleaved species (Tilia amurensis) in Changbai Mountains forests that were subjected to soil nitrogen and moisture treatments. RL increased as nitrogen deposition increased to a particular level (46 kg/ha/y N), when enzymes and mineral nutrients became imbalanced. RL decreased as soil moisture decreased to a drought level. Light inhibition degree behaved inversely to RL, and the average range for both species combined was 42.0%-78.4% for nitrogen treatments and 58.3%-87.0% for soil moisture treatments. The results are fundamental to accurate modeling of terrestrial carbon budgets and assessment of the carbon economy in forest ecosystems in a rapidly changing climate. [ABSTRACT FROM AUTHOR]

Soil respiration in forests contributes to significant carbon dioxide emissions from terrestrial ecosystems but it varies both spatially and seasonally. Both abiotic and biotic factors influence soil respiration but their relative contribution to spatial and seasonal variability remains poorly understood, which leads to uncertainty in models of global C cycling and predictions of future climate change. Here, we hypothesize that tree diversity, soil diversity, and soil properties contribute to local-scale variability of soil respiration but their relative importance changes in different seasons. To test our hypothesis, we conducted seasonal soil respiration measurements along a local-scale environmental gradient in a temperate forest in Northeast China, analyzed spatial variability of soil respiration and tested the relationships between soil respiration and a variety of abiotic and biotic factors including topography, soil chemical properties, and plant and soil diversity. We found that soil respiration varied substantially across the study site, with spatial coefficients of variation (CV) of 29.1%, 27.3% and 30.8% in spring, summer, and autumn, respectively. Soil respiration was consistently lower at high soil water content, but the influence of other factors was seasonal. In spring, soil respiration increased with tree diversity and biomass but decreased with soil fungal diversity. In summer, soil respiration increased with soil temperature, whereas in autumn, soil respiration increased with tree diversity but decreased with increasing soil nutrient content. However, soil nutrient content indirectly enhanced soil respiration via its effect on tree diversity across seasons, and forest stand structure indirectly enhanced soil respiration via tree diversity in spring. Our results highlight that substantial differences in soil respiration at local scales was jointly explained by soil properties (soil water content and soil nutrients), tree diversity, and soil fungal diversity but the relative importance of these drivers varied seasonally in our temperate forest. [ABSTRACT FROM AUTHOR]

Soil respiration (Rs) consists of autotrophic (Ra) and heterotrophic (Rh) respiration, and the metabolic responses of Ra and Rh are supposedly affected by environmental factors. Our hypothesis was that the contribution of Ra and Rh would be affected by seasons. To characterize seasonal patterns of Rs, sap flow (as an indicator of photosynthetic activity), and environmental factors, we continuously measured temporal variation in Rs using an automated opening and closing chamber system in a cool temperate forest. Rs had counterclockwise hysteresis (Rs: spring < autumn) with soil temperature at a depth of 5 cm. Daily maximum Rs had a significant positive relationship with daily maximum sap flow over three seasons, and daily maximum sap flow was lower in autumn than in spring (Ra: spring > autumn). The amount of leaf litterfall increased significantly from August to October; the high Rs in autumn would be due to an increase in Rh (Rh: spring < autumn). These results suggest that Ra contributes more in spring than in autumn, and the contribution of Rh is high from summer through autumn. [ABSTRACT FROM AUTHOR]

• Nonlinear thresholds of quick flow were related to the sum of ASW and rainfall. • A fast increase in quick flow was observed above the rising threshold in both forests. • Broadleaf forest showed faster increase in quick flow above the rising threshold phase. • Soil thickness expanded soil water storage deficit and disrupted flow path connectivity. The nonlinear behavior of hillslope quick flow (QF) in response to variations in soil thickness and forest types remains unclear due to uncertainties or unknown features in field studies, hindering accurate interpretation of their impact on QF. To this end, three artificial physical models with different mineral soil thickness (0 cm, shallow 50 cm, and thick 150 cm) beneath a 50 cm root zone for each forest (i.e., the broadleaf and coniferous forests) were constructed and intensively monitored. Results showed that the threshold-based QF was a function of the sum of rainfall amount and antecedent soil wetness, which was isolated by two breakpoints—the generation and rising thresholds—across all artificial physical models. These breakpoints indicated the storage thresholds for QF generation and the transition from slow to fast QF in humid forests. The broadleaf forest required more rainfall to reach the generation threshold due to higher litter interception capacity and soil water deficit by plant water uptake compared to the coniferous forest. Above the rising threshold phase, QF increased rapidly and linearly due to the establishment of flow path connectivity driven by soil water content, with the broadleaf forest showing a faster increase in QF than the coniferous forest. The dominant preferential/unsaturated flow in humid mountainous areas was also confirmed. Preferential flow paths were well-connected in the root zone and extended to the shallow mineral soil, resulting in an insignificant difference in mean QF between forest types. The thick mineral soil disrupted preferential flow path connectivity, leading to a greater amount of water being laterally diverted and retained within the soil. Consequently, a sharp decrease in QF was observed in hillslope with the root zone capping thick mineral soil for both forests. Our results highlighted the application of threshold-based theory and controlled experiments in enhancing the comprehension of the fundamental physical mechanism governing runoff processes in humid forests. [ABSTRACT FROM AUTHOR]

Surface topography can influence flow pathways and the location of runoff source areas and water transport in steep headwater catchments. However, the influence of topography on spatial patterns of residual soil moisture is less well understood. We measured soil volumetric water content (VWC) on 14 dates at 0–30 and 30–60 cm depth at 54 sites on a steep, 10 ha north‐facing forested slope in the west‐central Cascades Mountains of Oregon, USA. Spatial patterns in VWC were persistent over time, and contrary to expectations VWC at 30–60 cm depth was greater on divergent than convergent slopes, especially during wet periods (R2 = 0.27, p < 0.001). Vegetation characteristics were assessed for all VWC monitoring locations and soil properties were determined for 13 locations as local factors that affect spatial patterns in VWC. Mean VWC over all dates was negatively correlated to gravimetric rock content (R2 = 0.28, p = 0.03) and positively correlated to water storage at field capacity (R2 = 0.56, p < 0.01). The variability in rock content in quick‐draining soils influenced soil‐water retention, and by extension, created spatially heterogenous but temporally persistent patterns in VWC. While spatial patterns were persistent, they were not easily explained by surficial topography in a steep, mountainous landscape with rocky, well‐drained soils. Further research is needed to understand if combined soil‐terrain metrics would be a more useful proxy for VWC than terrain‐based wetness metrics alone. Plain Language Summary: Surface topography can direct the movement of water through the subsurface in ways that affect water availability for plants. When soils saturate, excess water may move downslope, leading to more water in hillslope hollows. However, in places where soils do not become saturated, the factors controlling the spatial patterns in water availability are less well understood. We designed our study to test if the redistribution of water, following surface topography, controlled the spatial patterns of soil moisture across a steep forested hillslope with well‐drained soils. We found that spatial patterns of soil moisture were similar throughout the year, but the movement of water downslope was not responsible for these patterns because soils were drier in hillslope hollows and wetter on the intervening ribs. Additionally, we found that the amount of water in the soil was strongly related to rock content. Thus, the presence of rocks may contribute to the persistence in spatial patterns of soil moisture. We recommend future research focuses on understanding how spatial patterns in forest soil properties affect soil water storage. This will allow us to improve computer models that predict water stress in forest ecosystems, and thereby, help forest managers develop plans to mitigate water stress. Key Points: Terrain‐based metrics were poor predictors of shallow soil water contentSoil properties were strong predictors of shallow soil water contentSpatial patterns in soil water content were persistent across seasonal variations in catchment wetness state [ABSTRACT FROM AUTHOR]

The purpose of this paper is to test the hypothesis that the natural soil compaction and its related soil properties in forests are significantly different from the soil compaction in arable soils, under the same climate, relief, geology and soil conditions. This study was carried out in the southern part of the Romanian Plain, during the summer of 2018. The soil contains relatively deep fertile topsoil followed by subsoil enriched in clay. The used methodology consists of digging four pairs of 1.5 m depth soil profiles performed in both forested and arable land, from where both disturbed and undisturbed soil samples were taken and analyzed in the lab. The different environmental conditions and human activities produced modifications in soil physical properties. The results emphasize significant differences between bulk density (BD), other soil physical properties and various soil water content values for the soil matric potential of pF 0, 2.52 and 4.2 in the forest soils, versus the arable soils. The upper part of the soil profile is usually modified in the arable land and soil physical properties worsen. BD was as much as 1.6 kg dm−3 in the topsoil of the arable land, being by 25% higher versus the forested land. However, in subsoil BD is significantly higher in the oak tree forest soils versus the arable soils probably due to the pressure exerted by the permanent thicker tree roots. The potential impact of this research should be regarded in the context of soil management, emphasizing the differences between these two land uses. Thus, for an improvement of the physical properties of arable topsoil and to increase soil carbon sinks by reducing CO2 deposition in the atmosphere, afforestation for limited periods might be considered. [ABSTRACT FROM AUTHOR]

In order to demonstrate the response of soil microbial community and hydrothermal environment to nitrogen deposition (low nitrogen N20: 20 kg.hm-2; medium nitrogen N40: 40 kg.hm-2; high nitrogen N60: 60 kg.hm-2, and contrast N0: 0 kg.hm-2), a Pinus massoniana forest in Central Asia was chosen to do the nitrogen deposition simulation experiment. This research is aimed to provide a theoretical evidence for the protection of soil ecosystem under different forest types in china. The results showed that: soil microbes of Pinus massoniana forest were in seasonal changes (spring, autumn, winter, and summer). Differences in different seasons were significant: a very significant quadratic relationship was shown between soil microbes and soil temperature. However, the relationship between soil microbes and the soil water content was not closely related. The N deposition reduced the relationship between microbes and temperature but increased the correlation between microbes and water content: effects of N deposition on soil temperature and soil water content were significant, but the effects were in small scale, and the effects of N deposition on soil microbial community structure were significant. [ABSTRACT FROM AUTHOR]

In tropical rainforests the pattern and quantity of rainfall influence various ecosystem functions, as well as forest stand structure by controlling soil water status. In the present study, spatial distribution of volumetric soil water content (VSWC), measured over 1 year, was compared with stand structure in a Southeast Asian tropical rainforest. Forest surface-layer VSWC was determined by the relative elevation and soil physical properties. The patterns of spatial variation in surface-layer VSWC and residual VSWC during dry period suggested that drier surface soil areas had developed forest soil texture with larger pore size. There were no relationships between VSWC and tree mortality or number of trees for all soil layer, but a significant negative relationship was found between surface-layer VSWC and basal area, and thus also biomass. This could be due to the preference of trees for drier surface areas with larger median pore size. The positive relationship between mean diameter at breast height (dbh) and deep layer soil water content might indicate the preference of deep layer water by big trees. On the other hand, the number of trees in Dipterocarpaceae, as well as of emergent species trees, was negatively correlated with SWC in deep soil layer. [ABSTRACT FROM AUTHOR]

The presented study shows the results of measuring soil respiration in typical burozems (Dystric Cambisols) under mixed Korean pine-broadleaved forests in the southern part of the Primorskii (Far East) region of Russia growing under conditions of monsoon climate. The measurements were performed in 2014-2016 by the chamber method with the use of a portable infrared gas analyzer. Relative and total values of the CO2 efflux from the soil surface on four model plots were determined. The intensity of summer emission varied from 2.25 to 10.97 μmol/(m2 s), and the total CO2 efflux from the soils of four plots varied from 18.84 to 25.56 mol/m2. It is shown that a larger part of seasonal variability in the soil respiration is controlled by the soil temperature (R2 = 0.5-0.7); the soil water content also has a significant influence on the CO2 emission determining about 10% of its temporal variability. The daily dynamics of soil respiration under the old-age (200 yrs) forest have a significant relationship with the soil temperature (R2 = 0.51). The pyrogenic transformation of Pinus koraiensis forests into low-value oak forests is accompanied by an increase in the СО2 efflux from the soil. [ABSTRACT FROM AUTHOR]

• Increasing woodland use intensity (WUI) had significant impacts on soil EMF. • WUI can indirectly affect soil EMF through microbial diversity and network complexity. • Soil EMF was regulated by soil microbial diversity and network complexity. • Soil water content mainly affected EMF both in the dry and wet seasons. Global change (e.g., woodland use change) affects ecosystem functioning directly by modifying various physicochemical processes and indirectly by altering biological metabolism and interactions. To determine whether woodland use intensity (WUI) affects soil ecosystem multifunctionality (EMF) via microbial diversity and network complexity and whether the dry and wet season drivers are the same. We analyzed the relationships of soil EMF to soil microbial (bacteria and fungi) diversity and network complexity, considering the WUI in the monsoon evergreen broad-leaved forest in dry and wet seasons in subtropical China. The findings showed that with increasing of WUI, soil microbial (bacteria and fungi) diversity increased, whereas soil EMF decreased. The microbial network complexity decreased in the dry season but increased in the wet season. Soil EMF was jointly regulated by indirect effects from WUI in addition to microbial diversity and network complexity. WUI and abiotic factors can indirectly affect soil EMF through microbial diversity and network complexity, among which soil water content is the main environmental factor affecting soil EMF in dry season, and fungal diversity is the main factor affecting soil EMF in wet season. In addition to demonstrating how multiple anthropogenic drivers differ in their impact via different pathways on soil EMF components, our results will facilitate the prediction of ecosystem responses to multiple simultaneous environmental changes. [ABSTRACT FROM AUTHOR]

Trees in alley-cropping systems (AC) were reported to develop deeper fine roots compared to forest trees and that they can modify soil water (SWC), mineral nitrogen (SMN) and organic matter (SOM) content. However, intercropping young trees has not been studied extensively. This study aimed to count tree fine roots abundance (TFRA) along a chronosequence of AC stands, to determine factors explaining its variability and to highlight its effects on soil resources. Seventeen alley-cropping plots ranging from 3 to 12 years old were chosen on farms in northern France. TFRA was measured by the core break method using soil samples collected at 0, 1, 3 and 10 m from a referent tree (a maple, a hybrid walnut or a hornbeam) down to 2 m depth. Before four years old, tree fine roots colonized the topsoil (0–30 cm) in rows and then mainly grew vertically from 4 to 6 years old, before laterally exploring deep soil layer (1–2 m) beyond this age. Stepwise analyses showed that stand age, tillage frequency and crop rotation duration explained 60 % of the variability of the sum of TFRA calculated for all soil layers at all distances from the tree row. The SWC was negatively correlated to TFRA suggesting that as trees get older, they dried the deep soil layer below the crop rooting zone and increased the soil depth able to store autumn and winter rainfall. No significant effect of either stand age or distance from tree rows was observed for SMN. It varied significantly with soil depth (R² = 0.3***) and was strongly correlated with soil nitrate content (R² = 0.97***). The soil ammonium content was significantly correlated with TFRA, suggesting that tree fine roots favor ammonium production or accumulation in soil, which may potentially allow for a reduction in the mineral nitrogen (N) mobility for leaching. Finally, we found a significantly high SOM correlated with TFRA only in topsoil on the tree rows at our oldest stands. No change of SOM was observed in the deep soil layer regardless of stand age. From this study, we concluded that fine root plasticity of intercropped trees occurred at early stage and may contribute with age to a better use of soil water, to managing the soil mineral N dynamic and to sequestrating carbon, at least in tree rows. • Tree fine roots abundance was quantified along a chronosequence of young agroforestry stands ranging from 3 to 12 years old. • Intercropped trees developed vertical fine roots on rows from 4 to 6 years old before laterally exploring deep soil layer thanks to soil tillage and crop rotation duration. • Tree fine roots dried the deep soil layer and promote soil ammonium accumulation in tree rows. • A significant increase in soil organic matter content was observed in tree rows of 10–12 years old partially due to tree fine roots abundance. [ABSTRACT FROM AUTHOR]

Soil CO2 efflux ( Fsoil) is the largest source of carbon from forests and reflects primary productivity as well as how carbon is allocated within forest ecosystems. Through early stages of stand development, both elevated [ CO2] and availability of soil nitrogen (N; sum of mineralization, deposition, and fixation) have been shown to increase gross primary productivity, but the long-term effects of these factors on Fsoil are less clear. Expanding on previous studies at the Duke Free-Air CO2 Enrichment ( FACE) site, we quantified the effects of elevated [ CO2] and N fertilization on Fsoil using daily measurements from automated chambers over 10 years. Consistent with previous results, compared to ambient unfertilized plots, annual Fsoil increased under elevated [ CO2] (ca. 17%) and decreased with N (ca. 21%). N fertilization under elevated [ CO2] reduced Fsoil to values similar to untreated plots. Over the study period, base respiration rates increased with leaf productivity, but declined after productivity saturated. Despite treatment-induced differences in aboveground biomass, soil temperature and water content were similar among treatments. Interannually, low soil water content decreased annual Fsoil from potential values - estimated based on temperature alone assuming nonlimiting soil water content - by ca. 0.7% per 1.0% reduction in relative extractable water. This effect was only slightly ameliorated by elevated [ CO2]. Variability in soil N availability among plots accounted for the spatial variability in Fsoil, showing a decrease of ca. 114 g C m−2 yr−1 per 1 g m−2 increase in soil N availability, with consistently higher Fsoil in elevated [ CO2] plots ca. 127 g C per 100 ppm [ CO2] over the +200 ppm enrichment. Altogether, reflecting increased belowground carbon partitioning in response to greater plant nutritional needs, the effects of elevated [ CO2] and N fertilization on Fsoil in this stand are sustained beyond the early stages of stand development and through stabilization of annual foliage production. [ABSTRACT FROM AUTHOR]

Abstract: Forest soils contain the largest carbon stock of all terrestrial biomes and are probably the most important source of carbon dioxide (CO2) to atmosphere. Soil CO2 fluxes from 54 to 72-year-old monospecific stands in Rwanda were quantified from March 2006 to December 2007. The influences of soil temperature, soil water content, soil carbon (C) and nitrogen (N) stocks, soil pH, and stand characteristics on soil CO2 flux were investigated. The mean annual soil CO2 flux was highest under Eucalyptus saligna (3.92μmol m−2 s−1) and lowest under Entandrophragma excelsum (3.13μmol m−2 s−1). The seasonal variation in soil CO2 flux from all stands followed the same trend and was highest in rainy seasons and lowest in dry seasons. Soil CO2 flux was mainly correlated to soil water content (R 2 =0.36–0.77), stand age (R 2 =0.45), soil C stock (R 2 =0.33), basal area (R 2 =0.21), and soil temperature (R 2 =0.06–0.17). The results contribute to the understanding of factors that influence soil CO2 flux in monocultural plantations grown under the same microclimatic and soil conditions. The results can be used to construct models that predict soil CO2 emissions in the tropics. [Copyright &y& Elsevier]

Forest soil is a huge reserve of carbon in the biosphere. Therefore to understand the carbon cycle in forest ecosystems, it is important to determine the dynamics of soil CO2 efflux. This study was conducted to describe temporal variations in soil CO2 efflux and identify the environmental factors that affect it. We measured soil CO2 efflux continuously in a beech secondary forest in the Appi Highlands in Iwate Prefecture for two years (except when there was snow cover) using four dynamic closed chambers that automatically open after taking measurements. Temporal changes in soil temperature and volumetric soil water content were also measured at a depth of 5 cm. The soil CO2 efflux ranged from 14 mg CO2 m−2 h−1 to 2,329 mg CO2 m−2 h−1, the peak occurring at the beginning of August. The relationship between soil temperature and soil CO2 efflux was well represented by an exponential function. Most of temporal variation in soil CO2 efflux was explained by soil temperature rather than volumetric soil water content. The Q 10 values were 3.7 ± 0.8 and estimated annual carbon emissions were 837 ± 210 g C m−2 year−1. These results provide a foundation for further development of models for prediction of soil CO2 efflux driven by environmental factors. [ABSTRACT FROM AUTHOR]

Abstract: Our objectives were to determine both spatial and temporal variations in soil respiration of a mixed deciduous forest, with soils exhibiting contrasting levels of hydromorphy. Soil respiration (R S) showed a clear seasonal trend that reflected those of soil temperature (T S) and soil water content (W S), especially during summer drought. Using a bivariate model (RMSE=1.03), both optimal soil water content for soil respiration (W SO) and soil respiration at both 10°C and optimal soil water content (R S10) varied among plots, ranging, respectively, from 0.25 to 0.40 and from 2.30 to 3.60μmolm−2 s−1. Spatial variation in W SO was related to bulk density and to topsoil N content, while spatial variation in R S10 was related to basal area and the difference in pH measured in water or KCl suspensions. These results offer promising perspectives for spatializing ecosystem carbon budget at the regional scale. [Copyright &y& Elsevier]

Land use, soil organic carbon content and bulk density are important factors influencing the occurrence of soil water repellency (SWR) and water flow through the soil profile. This study elucidated the relations between the repellency index (RI) and contents of soil organic carbon (SOC), organic-complexed aluminium (Al p) and iron (Fe p), pH, specific surface area (SSA) and bulk density in the ortstein horizon and the overlying and underlying horizons in two podzolic soils under grassland and forest. The soil under grassland was originally used as a cropland. The RI was quantified from the ethanol/water sorptivity ratio. The results showed that the RI in the ortstein horizon (20.6–27.3) in the air-dried state in both soils was many times greater than in the overlying (7.2–19.0) and underlying (6.1–11.5) horizons, whereas in moist conditions (from soil water content at sampling and air-dried state), all horizons were not repellent. The RI in the air-dried soil in the overlying horizon was higher under grassland than forest, whereas the inverse was true in both the ortstein and underlying horizons. SOC, organic-complexed Al p and Fe p contents, and SSA were the highest in the ortstein horizons. The pH value increased from 5.3 in the overlying horizon to 6.0 in the underlying horizon in the grassland soil. The corresponding values in the forest soil were 4.2 and 5.0. The RI values in the dry condition were significantly correlated with Al p (R2 = 0.976), Fe p (R2 = 0.798), SSA (R2 = 0.724), and SOC (R2 = 0.638). The study provides new information about the vertical distribution of soil water repellency in ortsteinic soils depending on the soil water status and the type of land use. The potential mechanisms of the distribution of SWR are discussed. • Ortstein and over and underlying horizons in moist condition were not water repellent. • In the air dried state ortstein horizon was more repellent than neighbouring horizons. • Soil water repellency in the air-dried state was correlated with Al, Fe, SSA and SOC. [ABSTRACT FROM AUTHOR]

• The effect of soil water content (SWC) on forest ecosystem water use efficiency (WUE) is evaluated. • WUE decreases with increasing SWC when vapor pressure deficit (VPD) is high. • WUE is independent on SWC when VPD is low. • The decrease of WUE with SWC at high VPD is due to increasing evaporation. Ecosystem water use efficiency (WUE), defined as the ratio of gross primary production (GPP) over evapotranspiration (ET), is a crucial variable that describes the tradeoff between ecosystem carbon uptake and water loss. However, the effects of soil moisture on WUE remain unclear. Based on hourly observations from 36 forest eddy-covariance sites globally and two percentile-based statistical models, we investigate the effects of soil water supply (i.e., volumetric soil water content, SWC) and vapor pressure deficit (VPD) on carbon-water coupling at the hourly timescale. We further decompose WUE into two components, i.e., the ratio between GPP and plant transpiration (T), WUE t (GPP/T), and the ratio of plant transpiration to total evapotranspiration, T/ET. Results show that the sensitivities of WUE to SWC significantly increased with the increase of VPD. The sensitivities of WUE to SWC are high at high VPD, due to a rapid decrease of T/ET with SWC at high VPD. At low VPD, WUE is largely independent of SWC. The contribution linked to WUE t is much weaker and no significant relationships are observed between the sensitivities of WUE t to SWC and VPD. Our study suggests that the change of ecosystem WUE with SWC is largely due to changes in T/ET, while WUE t remains rather constant. [ABSTRACT FROM AUTHOR]

• None of the published pedotransfer functions predicted soil water content satisfactorily. • New functions were developed with cropping-practice predictors to improve prediction of soil water content. • Random forests fit the observed data best, but predictions worsened during cross validation. • Multiple linear regression gave stable predictions of soil water content during both adjustment and cross-validation. Conservation agriculture has been developed as a means to improve the sustainability of agricultural systems and reduce drawbacks of conventional agricultural practices. Cropping practices can have a large influence on soil properties such as water retention. Proper tools are needed to assess and model effects of conservation agriculture on soil properties. As measuring soil water retention is expensive and time consuming, pedotransfer functions (PTFs) are now commonly used to predict them. The objectives of this study were to (i) present a new dataset of conservation agriculture data, (ii) assess performances of existing PTFs in predicting soil water retention of soils under conservation agriculture and (iii) develop new specific PTFs to predict water retention in conservation agriculture more accurately. We used data collected only in fields under conservation agriculture in France to evaluate several published PTFs with three evaluation criteria (RMSE, prediction bias (ME) and Nash-Sutcliffe Efficiency (EF)). We then developed new PTFs using three methods ― multiple linear regression, regression tree and random forest ― to predict soil water content at matric heads of -100 (θ 100 , field capacity for sandy soils), -330 (θ 330 , field capacity for other soils) and -15 000 cm (θ 15 000 , wilting point). Soil tillage, presence of a cover crop, rotation length and previous reduced/no tillage were used as predictors in addition to basic soil properties for regression trees and random forests. The quality of prediction (RMSE, ME and EF) was calculated for each new PTF using a cross-validation procedure. Generally, predictions of wilting point had lower absolute error than those of sandy-soil field capacity (RMSE = 0.044 and 0.066 cm3/cm3, respectively). EF was usually negative for all water contents. The cross-validation performance of the new PTFs was similar for multiple linear regression (RMSE: 0.028, ME: 0.000, EF: 0.34 for θ 100) and random forest (RMSE: 0.027, ME: 0.000, EF: 0.36 for θ 100), and generally worse for regression tree (especially EF). Multiple linear regression that did not consider cropping practices performed as well as random forest and thus did not identify any major influence of agricultural management on predicted water content. Future research on developing PTFs should focus on identifying more relevant predictors. [ABSTRACT FROM AUTHOR]