Global Sensitivity Analysis of the Advanced ORYZA-N Model with Different Rice Types and Irrigation Regimes.

Identifying important parameters in crop models is critical for model application. This study conducted a sensitivity analysis of 23 selected parameters of the advanced rice model ORYZA-N using the Extended FAST method. The sensitivity analysis was applied for three rice types (single-season rice in cold regions and double-season rice (early rice and late rice) in subtropical regions) and two irrigation regimes (traditional flood irrigation (TFI) and shallow-wet irrigation (SWI)). This study analyzed the parameter sensitivity of six crop growth outputs at four developmental stages and yields. Furthermore, we compared the variation in parameter sensitivity on model outputs between TFI and SWI scenarios for single-season rice, early rice, and late rice. Results indicated that parameters RGRLMX, FRPAR, and FLV0.5 significantly affected all model outputs and varied over developmental stages. Water stress in paddy fields caused by water-saving irrigation had more pronounced effects on single-season rice than on double-season rice.

Lignite bioorganic fertilizer enhanced microbial co-occurrence network stability and plant-microbe interactions in saline-sodic soil.

Lignite-converted bioorganic fertilizer substantially improves soil physiochemical properties, but little is known about how lignite bioorganic fertilizer (LBF) affects soil microbial communities and how the changed microbial communities impact their stability, functions, and crop growth in saline-sodic soil. Therefore, a two-year field experiment was conducted in saline-sodic soil in the upper Yellow River basin, Northwest China. Three treatments, i.e., the control treatment without organic fertilizer (CK), the farmyard manure treatment (FYM) amended with 21 t ha-1 (same as local farmers) sheep manure, and the LBF treatment amended with the optimal rate of LBF (3.0 and 4.5 t ha-1), were designed in this study. The results showed that after two years of application of LBF and FYM, the percentage of aggregate destruction (PAD) was significantly reduced by 14.4 % and 9.4 %, respectively, while the saturated hydraulic conductivity (Ks) was obviously increased by 114.4 % and 99.7 %, respectively. The LBF treatment significantly increased the contributions of nestedness to total dissimilarity by 101.4 % and 156.2 % in bacterial and fungal communities, respectively. LBF contributed to the shift from stochasticity to variable selection in the assembly of the fungal community. The LBF treatment enriched the bacterial classes of Gammaproteobacteria, Gemmatimonadetes, and Methylomirabilia and fungal classes of Glomeromycetes and GS13, which were mainly driven by PAD and Ks. Additionally, the LBF treatment significantly increased the robustness and positive cohesions and decreased the vulnerability of the bacterial co-occurrence networks in both 2019 and 2020 in comparison with the CK treatment, indicating that the LBF treatment increased stability of bacterial community. The relative abundance of chemoheterotrophy and arbuscular mycorrhizae in the LBF treatment were 89.6 % and 854.4 % higher than those in the CK treatment, respectively, showing that the LBF enhanced sunflower-microbe interactions. The FYM treatment improved the functions mainly regarding sulfur respiration and hydrocarbon degradation by 309.7 % and 212.8 % in comparison with the CK treatment, respectively. The core rhizomicrobiomes in the LBF treatment showed strong positive connections with the stabilities of both bacterial and fungal co-occurrence networks, as well as the relative abundance and potential functions of chemoheterotrophy and arbuscular mycorrhizae. These factors were also linked to the growth of sunflowers. This study reveals that the LBF improved sunflower growth due to enhance microbial community stability and sunflower-microbe interactions through altering core rhizomicrobiomes in saline-sodic farmland.

An improved estimation of soil water and salt dynamics by considering soil bulk density changes under freeze/thaw conditions in arid areas with shallow groundwater tables.

Soil bulk density (BD) is a parameter dependent on soil texture, compositions of soil minerals and organic matter and the extent of soil compaction. Seasonal freeze/thaw in arid areas with shallow groundwater tables (AASGT) may significantly change BD and hence soil hydrothermal properties and water holding capacity. Therefore, quantifying soil bulk density changes (BDC) under freeze/thaw conditions can improve estimates of soil water-salt dynamics in AASGT. In this study, we conducted field experiments to investigate the soil water-salt dynamics under freeze/thaw conditions from three typical land-use types (i.e., farmland, woodland, and natural land) in the upper Yellow River basin, China. We proposed a method to estimate BDC, which can better describe the soil water-salt dynamics during the freeze/thaw period. Our results showed marked BDC occurred in all layers within the 0-100 cm profile in natural land, while mainly at the 20-80 cm profile in farmland. During the freezing period, BD in farmland and natural land first decreased rapidly and then remained relatively stable until the thawing period started. After that, BD gradually increased during the thawing period. The largest BDC in farmland and natural land were 0.48 g cm-3 (occurring at the 30-40 cm layer) and 0.43 g cm-3 (occurring at the 80-90 cm layer), respectively, close to 30 % of their initial values. The differences in BDC between the three land-use types were mainly owing to their differences in groundwater table depth, initial soil salt concentration, soil texture, and surface coverage conditions. Moreover, in farmland and natural land, ignoring BDC resulted in different degrees of overestimation or underestimation in soil water content, water fluxes, and soil hydrothermal properties in the selected soil layers. This study demonstrates that considering BDC can improve the accuracy of soil water-salt dynamics estimation in AASGT under freeze/thaw conditions.

Modeling the effects of temperature on the migration and transformation of nitrate during riverbank filtration using HYDRUS-2D.

Riverbank filtration is a natural aquifer-based process. The nitrogen dynamics in a riverbank filtration system are affected by many factors, including temperature, water quality, and travel time, which cannot be quantified easily. In this study, a field experiment was conducted to investigate nitrogen transport during riverbank filtration. The HYDRUS-2D software package was used to investigate and quantify the factors that affect the fate of nitrogen. The effects of temperature, water quality, and travel time on nitrate transport were considered. The model was calibrated and validated using field experimental data from the river water and groundwater during riverbank filtration at different periods. The results showed that HYDRUS-2D adequately simulated nitrate transport during riverbank filtration. The denitrification rate constant exhibited a positive exponential relationship with temperature. An empirical formula describing this relationship in riverbank filtration was developed and validated. In addition, the denitrification rate can be quantified within a specified temperature data range under field conditions. Compared with indoor experimental conditions, for the same temperature, there was a 10-fold increase in the denitrification rate constant under field conditions. The results showed that most of the nitrate removal occurred in the riparian zone at high temperatures during riverbank filtration. We concluded that the fate of nitrate in the riparian zone is strongly controlled by groundwater temperature. Travel time also plays an important role in nitrate removal during riverbank filtration.

Experimental investigation and simulation of nitrogen transport in a subsurface infiltration system under saturated and unsaturated conditions.

The nitrogen dynamics in a subsurface infiltration system (SIS) are affected by many factors, including temperature, system design, and feed water quality, which are not easily quantified. In this study, a column experiment was conducted to simulate an SIS. The HYDRUS-1D software package was used to investigate and quantify the factors that affect nitrate transport in an SIS. Three treatments were carried out based on different hydraulic conditions, including continuous wetting (CW), wetting/drying (WD), and a specific hydraulic loading rate (SH). The effects of hydraulic conditions and temperature on nitrate transformation were investigated. The model was calibrated and validated using two-year experimental data. Simulations of cumulative outflow volume and nitrate concentration fitted well with the observations. Among the three SISs, the denitrification rate was greatest under unsaturated conditions at high water temperature. The denitrification rate constant had an exponential relationship with temperature. An empirical formula describing this relationship was developed and validated in the SIS. The results showed that the SH column attained the greatest nitrate removal efficiency, mainly due to its low hydraulic loading and long retention time. Overall, the results showed that HYDRUS-1D adequately simulated nitrate transport through the soil column under different temperature and hydraulic conditions in an SIS. The fate of nitrate was directly controlled by the water temperature and hydraulic conditions.

Responses of crop growth and water productivity to climate change and agricultural water-saving in arid region.

Climate change and associated elevated atmospheric CO2 concentration and rising temperature have become a great challenge to agricultural production especially in arid and semiarid regions, and a great concern to scientists worldwide. Thus, it is very important to assess the response of crop growth and water productivity to climate change projections, which in turn can help devise adaptive strategies to mitigate their impact. An agro-hydrological model with well consideration of CO2 effects on both the stomatal conductance and leaf area was established. The model was well calibrated and validated using the data collected from the middle oasis of Heihe River basin, northwest China, which was selected as a typical arid region. Simulations of soil water contents and crop growth matched well with observations. Then various scenarios were designed with considering three climate change alternatives (RCP 2.6, RCP 4.5 and RCP 8.5) and three agricultural water-saving alternatives in the context of irrigation water availability being constant. Responses of crop growth and water productivity were predicted for thirty years from 2018 to 2047. As compared to current situation, there would be a reduction of 3.4-8.6% in crop yield during the period of 2018-2027 and an increase of 1.5-18.7% in crop yield during the period of 2028-2047 for seed corn, and an increase of 7.4-26.7% in crop yield during the period of 2018-2047 for spring wheat, respectively. Moreover, results showed an increase in water productivity ranged from 14.3% to 44.5% for seed corn and from 34.7% to 52.0% for spring wheat, respectively. Furthermore, adaptive strategies to climate change were recommended for the seed corn and spring wheat, respectively. Our results are expected to provide implications for devising adaptive strategies to changing environments in other arid and irrigation-fed areas.

Modeling solute transport in one-dimensional homogeneous and heterogeneous soil columns with continuous time random walk.

In this paper, we used the continuous time random walk (CTRW) framework to characterize the transport process in 1250-cm long one-dimensional homogenous and heterogeneous soil columns at the experiments conducted by Huang et al. [Huang, K., Toride, N., van Genuchten, M.Th., 1995. Experimental investigation of solute transport in large, homogeneous and heterogeneous, saturated soil columns. Trans. Porous Media. 18, 283-302]. The transport process was also simulated by using the advection-dispersion equation (ADE) and the spatial fractional advection-dispersion equation (FADE) for comparison. In the homogeneous soil column, the non-Fickian behavior is found at the distances less than 1000cm with beta values larger than 1.60, but less than 2, and Fickian form transport is obtained at distances larger than 1000cm with beta values larger than 2. In the heterogeneous soil column, we found the most anomalous behavior at distances from 200cm to 700cm with beta values ranging from 0.894 to 0.958, and non-Fickian transport process is observed at distances larger than 800cm with beta values in the range between 1 and 1.3. More significant non-Fickian behavior is found for transport in the heterogeneous soil column than that in the homogeneous soil column. The CTRW fits to the breakthrough curves (BTCs) have lower values of root mean square error (RMSE) and higher values of determination coefficient (r(2)), with respect to the fits of ADE and FADE. The CTRW model also is better captures the full evolution of BTCs, and especially their tails.

Evidence of one-dimensional scale-dependent fractional advection-dispersion.

A semi-analytical inverse method and the corresponding program FADEMain for parameter estimation of the fractional advection-dispersion equation (FADE) were developed in this paper. We have analyzed Huang et al.'s [Huang, K., Toride, N., van Genuchten, M.Th., 1995. Experimental investigation of solute transport in large homogeneous and heterogeneous saturated soil columns. Trans. Porous Media 18, 283-302.] laboratory experimental data of conservative solute transport in 12.5-m long homogeneous and heterogeneous soil columns to test the non-Fickian dispersion theory of FADE. The dispersion coefficient was calculated by fitting the analytical solution of FADE to the measured data at different transport scales. We found that the dispersion coefficient increased exponentially with transport scale for the homogeneous column, whereas it increased with transport scale in a power law function for the heterogeneous column. The scale effect of the dispersion coefficient in the heterogeneous soil was much more significant comparing to that in the homogeneous soil. The increasing rate of dispersion coefficient versus transport distance was smaller for FADE than that for the advection-dispersion equation (ADE). Finite difference numerical approximations of the scale-dependent FADE were established to interpret the experimental results. The numerical solutions were found to be adequate for predicting scale-dependent transport in the homogeneous column, while the prediction for the heterogeneous column was less satisfactory.