Improvement of multi-layer soil moisture prediction using support vector machines and ensemble Kalman filter coupled with remote sensing soil moisture datasets over an agriculture dominant basin in China

Soil moisture prediction is of great importance in crop yield forecasting and drought monitoring. In this study, the multi-layer root zone soil moisture (0-5, 0-10, 10-40 and 40-100 cm) prediction is conducted over an agriculture dominant basin, namely the Xiang River Basin, in southern China. The support vector machines (SVM) coupled with dual ensemble Kalman filter (EnKF) technique (SVM-EnKF) is compared with SVM for its potential capability to improve the efficiency of soil moisture prediction. Three remote sensing soil moisture products, namely SMAP, ASCAT and AMSR2, are evaluated for their performance in multi-layer soil moisture prediction with SVM and SVM-EnKF, respectively. Multiple cases are designed to investigate the performance of SVM, the effectiveness of coupling dual EnKF technique and the applicability of the remote sensing products in soil moisture prediction. The main results are as follows: (a) The efficiency of soil moisture prediction with SVM using meteorological variables as inputs is satisfactory for the surface layers (0-5 and 0-10 cm), while poor for the root zone layers (10-40 and 40-100 cm). Adding SMAP as input to SVM can improve its performance in soil moisture prediction, with more than 47% increase in the R-value and at least 11% reduction in RMSE for all layers. However, adding ASCAT or AMSR2 has no improvement for its performance. (b) Coupling dual EnKF can significantly improve the performance of SVM in the soil moisture prediction of both surface and the root zone layers. The increase in R-value is above 80%, while the reduction in BIAS and RMSE is respectively more than 90% and 63%. However, adding remote sensing soil moisture products as inputs can no further improve the performance of SVM-EnKF. (c) The SVM-EnKF can eliminate the influence of remote sensing soil moisture extreme values in soil moisture prediction, therefore, improve its accuracy.     

Sensitivity of hydrological variables to climate change in the Haihe River basin,China

Using the defined sensitivity index, the sensitivity of streamflow, evapotranspiration and soil moisture to climate change was investigated in four catchments in the Haihe River basin. Climate change contained three parts: annual precipitation and temperature change and the change of the percentage of precipitation in the flood season (P_f). With satisfying monthly streamflow simulation using the variable infiltration capacity model, the sensitivity was estimated by the change of simulated hydrological variables with hypothetical climatic scenarios and observed climatic data. The results indicated that (i) the sensitivity of streamflow would increase as precipitation or P_f increased but would decrease as temperature increased; (ii) the sensitivity of evapotranspiration and soil moisture would decrease as precipitation or temperature increased, but it to P_f varied in different catchments; and (iii) hydrological variables were more sensitive to precipitation, followed by P_f, and then temperature. The nonlinear response of streamflow, evapotranspiration and soil moisture to climate change could provide a reference for water resources planning and management under future climate change scenarios in the Haihe River basin. Copyright © 2011 John Wiley & Sons, Ltd.     

Using data‐driven methods to explore the predictability of surface soil moisture with FLUXNET site data

Soil moisture (SM) is a key variable of land surface‐atmosphere interactions. Data‐driven methods have been used to predict SM, but the predictability of SM has not been well evaluated. This study investigated what variables and methods can be used to better predict SM for leading times of 7 days or longer with a global coverage of FLUXNET site data for the first time. Three machine‐learning models, that is, Bayesian linear regression, random forest, and gradient boosting regression tree, are used for the prediction. Variables including atmospheric forcing, surface soil temperature, time variables (year, day of year, and hour), the Fourier transformation of time variables, and lagged SM (7‐ to 14‐day lagged) were sequentially added into models. A framework with five experiments is designed for factorial exploration of SM predictability. A stepwise method was used to build the best models for each site. The performance of regression models became better when adding more explaining variables in most cases. The results showed that from 50 to 95% of variation of the best models can be explained. The important explaining variables are lagged surface SM, followed by day of year, year, soil temperature, and atmospheric forcing. The predictability of SM depends highly on SM memory characteristics and the persistence of seasonality. The effect of SM memory characteristics on SM prediction as an initial condition question has been widely discussed in this paper. Our results also provide an insight that mechanisms of seasonality effects on SM should be also paid more attention to.     

基于土壤墒情抗差的降雨径流预报模型——设计与实验流域应用

本文旨在将实时监测得到的土壤墒情转化为流域水文模型可以直接使用的土壤含水量,论证将实时土壤墒情资料用于实时预报的可行性;利用实时监测土壤墒情,改进传统的模型结构,设计基于实测土壤墒情的降雨径流水文预报模型.采用土壤含水量误差抗差估计技术以抵御观测资料粗差的影响,提高系统的稳定性;并在此基础上提出了土壤含水量系统响应修正方法,以提高模型计算精度.将该模型应用于实验流域——宝盖洞流域进行应用检验,洪水模拟合格率达到92.3%,整体模拟精度达到甲级.     

On the role of soil moisture in daytime evolution of temperatures

This paper explores the relationship between temperature, evaporation and soil moisture using a planetary boundary layer (PBL) model. It focuses on illustrating and quantifying the effect of soil moisture on the evolution of daytime temperatures. A simple convective PBL model coupled with the Penman–Monteith (PM) equation is used to estimate evapotranspiration. Following calibration and sensitivity analysis, the model was used to simulate the relative impact of dry and wet soil moisture conditions on daytime temperatures by changing the surface resistance parameter in the PM equation. It was found that the maximum temperature that can be reached during a day is constrained by the amount of soil moisture and the available net radiation, confirming previously published results. Higher temperatures can be reached with greater net radiation and dry soil moisture conditions. Copyright © 2012 John Wiley & Sons, Ltd.     

The effects of initial soil moisture conditions on swale flow hydrographs

The effects of soil water content (SWC) on the formation of run‐off in grass swales draining into a storm sewer system were studied in two 30‐m test swales with trapezoidal cross sections. Swale 1 was built in a loamy fine‐sand soil, on a slope of 1.5%, and Swale 2 was built in a sandy loam soil, on a slope of 0.7%. In experimental runs, the swales were irrigated with 2 flow rates reproducing run‐off from block rainfalls with intensities approximately corresponding to 2‐month and 3‐year events. Run‐off experiments were conducted for initial SWC (SWC_ini) ranging from 0.18 to 0.43 m³/m³. For low SWC_ini, the run‐off volume was greatly reduced by up to 82%, but at high SWC_ini, the volume reduction was as low as 15%. The relative swale flow volume reductions decreased with increasing SWC_ini and, for the conditions studied, indicated a transition of the dominating swale functions from run‐off dissipation to conveyance. Run‐off flow peaks were reduced proportionally to the flow volume reductions, in the range from 4% to 55%. The swale outflow hydrograph lag times varied from 5 to 15 min, with the high values corresponding to low SWC_ini. Analysis of swale inflow/outflow hydrographs for high SWC_ini allowed estimations of the saturated hydraulic conductivities as 3.27 and 4.84 cm/hr in Swales 1 and 2, respectively. Such estimates differed from averages (N = 9) of double‐ring infiltrometer measurements (9.41 and 1.78 cm/hr). Irregularities in swale bottom slopes created bottom surface depression storage of 0.35 and 0.61 m³ for Swales 1 and 2, respectively, and functioned similarly as check berms contributing to run‐off attenuation. The experimental findings offer implications for drainage swale planning and design: (a) SWC_ini strongly affect swale functioning in run‐off dissipation and conveyance during the early phase of run‐off, which is particularly important for design storms and their antecedent moisture conditions, and (b) concerning the longevity of swale operation, Swale 1 remains fully functional even after almost 60 years of operation, as judged from its attractive appearance, good infiltration rates (3.27 cm/hr), and high flow capacity.     

Analysis of soil moisture dynamics beneath olive trees

Olive cultivation is a widespread land use in Mediterranean climates. The proper implementation of soil and water conservation practices in groves requires detailed knowledge of the governing hydrological processes. In this work topsoil moisture dynamics under wet and dry conditions and across a small catchment was investigated in the inter row (IR) and directly under the olive tree canopies (UC). We do this using a sensor network (11 stations) and a simple bucket model which was calibrated (June, 2011–2012) and validated (June, 2012–2013). During most of the year the normalized soil moisture contents (s) were greater in the IR than under UC, with an average normalized soil moisture difference of 0.12. The difference between UC and IR normalized soil moisture followed a seasonal pattern, reaching a maximum near 0.30 during spring. An analysis of the normalized soil moisture probability density functions (pdfs) was bimodal, showing characteristic dominant wet and dry soil moisture states, with the highest probability densities for the dry state. Overall the spatial variability of soil moisture was lower UC than in the IR. This was a result of the soil moisture buffering capacity of the canopy with respect to rainfall and evaporation, in addition to observed differences in soil properties. Hourly soil moisture data were successfully modelled (R² > 0.85), both UC and in the IR, yet with the inclusion of a simple formulation for canopy interception for the former. The results provide insight into how olive trees change hydrological processes in their neighbourhood. Copyright © 2016 John Wiley & Sons, Ltd.     

Spatial prediction of temporal soil moisture dynamics using HYDRUS‐1D

This article investigates the soil moisture dynamics within two catchments (Stanley and Krui) in the Goulburn River in NSW during a 3‐year period (2005–2007) using the HYDRUS‐1D soil water model. Sensitivity analyses indicated that soil type, and leaf area index were the key parameters affecting model performance. The model was satisfactorily calibrated on the Stanley microcatchment sites with a single point rainfall record from this microcatchment for both surface 30 cm and full‐profile soil moisture measurements. Good correlations were obtained between observed and simulated soil water storage when calibrations for one site were applied to the other sites. We extended the predictions of soil moisture to a larger spatial scale using the calibrated soil and vegetation parameters to the sites in the Krui catchment where soil moisture measurement sites were up to 30 km distant from Stanley. Similarly good results show that it is possible to use a calibrated soil moisture model with measurements at a single site to extrapolate the soil moisture to other sites for a catchment with an area of up to 1000 km² given similar soils and vegetation and local rainfall data. Site predictions were effectively improved by our simple data assimilation method using only a few sample data collected from the site. This article demonstrates the potential usefulness of continuous time, point‐scale soil moisture data (typical of that measured by permanently installed TDR probes) and simulations for predicting the soil wetness status over a catchment of significant size (up to 1000 km²). Copyright © 2012 John Wiley & Sons, Ltd.     

Summer soil moisture loss at Utah SNOTEL sites and streamflow recession at nearby gauges: variability in runoff generation and the potential for flow forecasting

This analysis compares decreases in soil moisture (SM) at Utah snow telemetry (SNOTEL) sites during the summer months with discharge at nearby stream gauging locations using data from water years 2008–2012. The following characteristics were evaluated: (1) the influence of the SM loss at mid‐depths (20 cm) on hydrograph recession, (2) the influence of moisture loss from deeper portions of the soil (50 cm) on late‐season baseflow and (3) the timing of this transition. Thirty‐four pairings were used between SNOTEL sites and nearby stream gauges in select locations throughout Utah, for 3–5 years each depending on data quality, to generate 143 total comparisons of soil moisture loss and stream discharge. Regressions were fairly strong (r² > 0.8) where the SNOTEL site was in a location with slow meltout rates, ample infiltration and minimal summer precipitation. In a few cases, the correlation was remarkably strong (r² > 0.95), even for SNOTEL sites located far from respective stream gauges (e.g. >30‐km, >1000‐m elevation difference for the best pairing). At such sites, transition timing in 2013 (between predominantly 20‐ vs 50‐cm SM loss) was well predicted from 2012 data given the similarity in water years, with discharges at the transition point less than 30% different than observed values in 2013. An index of the robustness of each pairing was generated to determine where this type of analysis might be most successful; however, results suggest that identification of high‐quality pairings may need to be site by site. Published 2015. This article is a U.S. Government work and is in the public domain in the USA.     

Modelling spatio-temporal soil moisture dynamics in mountain tundra

Soil moisture has a fundamental influence on the processes and functions of tundra ecosystems. Yet, the local dynamics of soil moisture are often ignored, due to the lack of fine resolution, spatially extensive data. In this study, we modelled soil moisture with two mechanistic models, SpaFHy (a catchment-scale hydrological model) and JSBACH (a global land surface model), and examined the results in comparison with extensive growing-season field measurements over a mountain tundra area in northwestern Finland. Our results show that soil moisture varies considerably in the study area and this variation creates a mosaic of moisture conditions, ranging from dry ridges (growing season average 12 VWC%, Volumetric Water Content) to water-logged mires (65 VWC%). The models, particularly SpaFHy, simulated temporal soil moisture dynamics reasonably well in parts of the landscape, but both underestimated the range of variation spatially and temporally. Soil properties and topography were important drivers of spatial variation in soil moisture dynamics. By testing the applicability of two mechanistic models to predict fine-scale spatial and temporal variability in soil moisture, this study paves the way towards understanding the functioning of tundra ecosystems under climate change.     

Sensitivity of soil moisture field evolution to rainfall forcing

In this paper the temporal behaviour of soil moisture is modelled and statistically characterized by use of the zero‐dimensional model for soil moisture dynamics and the rectangular pulses Poisson process model for rainfall forcing. The mean, covariance and spectral density function of soil moisture (both instantaneous and locally averaged cases) are analytically derived to evaluate its sensitivity to the model parameters. Finally, the probability density function of soil moisture is derived to evaluate the effect of rainfall forcing. All the model parameters used have been tuned to the Monsoon '90 data. Results can be summarized as follows. (1) Only the soil moisture model parameters (η and nZ_r) are found to affect the autocorrelation function in a distinguishable manner. On the other hand, both the rainfall model parameter (θ) and the effective soil depth (nZ_r) are found to be of impact to the soil moisture spectrum. However, as the smoothing (or damping) effect of soil is so dominant, about ±20% variation of one parameter seems not to affect significantly the second‐order statistics of soil moisture. (2) More difference can be found by applying a longer averaging time, which is found to obviously decrease the variance but increase the correlation even though no overlapping between neighbouring soil moisture data was allowed. (3) Among rainfall model parameters, the arrival rate (λ) was found to be most important for the soil moisture evolution. When increasing the arrival rate of rainfall, the histogram of soil moisture shifts its peak to a certain value as well as becomes more concentrated around the peak. However, by decreasing the arrival rate of rainfall, a much smaller (almost to zero) mean value of soil moisture was estimated, even though the total volume of rainfall remained constant. This indicates that desertification may take place without decreasing the total volume of rainfall. Copyright © 2005 John Wiley & Sons, Ltd.     

Investigating the impact of initial soil moisture conditions on total infiltration by using an adapted double-ring infiltrometer

ABSTRACT

This study presents an adaptation of the double-ring infiltrometer (DRI) device, which allows several infiltration experiments to be conducted at the same location. Hence, it becomes possible to use the DRI method to investigate infiltration behaviour under different initial soil moisture conditions. The main feature is the splitting of the inner ring into two parts. While the lower part remains in the soil throughout the investigation period, the upper part is attached to the lower one just before the infiltration experiment. This method was applied to eight test sites in an Alpine catchment, covering different land-use/cover types. The results demonstrated the applicability of the adapted system and showed correlations between total water infiltration and initial soil moisture conditions on pastures, independent of the underlying soil type. In contrast, no correlation was found at forest sites or wetlands. Thus, the study emphasizes the importance of paying special attention to the impact of initial soil moisture conditions on the infiltration—and consequently the runoff behaviour—at managed areas. Given the differences in the total infiltrated water of between 30 and 1306 mm, consideration of the interplay between initial soil moisture conditions, land-use/cover type, and soil properties in rainfall–runoff models is a prerequisite to predict runoff production accurately.

EDITOR Z.W. Kundzewicz; ASSOCIATE EDITOR not assigned     

Spatial and temporal dynamics of soil moisture after rainfall events along a slope in Regosols of southwest China

High frequency seasonal drought in purple soils (Regosols in FAO taxonomy) of the hilly upland areas of Sichuan basin, China, is one of the key restrictive factors for crop production. In order to manage irrigation and fertilizer application in these soils effectively, the soil water content in a sloped plot with 60 cm soil depth was measured by neutron probe devices to investigate the soil moisture regime during the 1998 rainy season after various amounts of rainfall events. The results showed that variation of soil moisture along the slope positions was highest in the top soil layer during the period of sporadic rainfall that did not induce any runoff. The coefficients of variation of soil moisture at various slope positions (upper, middle, and lower) are 17·36%, 8·95%, 10·25%, 8·58%, 8·05% and 9·21% at the 10 cm, 20 cm, 30 cm, 40 cm, 50 cm and 60 cm soil depths respectively. When surface runoff occurred, the soil moisture dynamics at various positions on the plot were then very different. Soil water content decreased more rapidly on the upper slope than on the middle and lower slope positions. When both surface runoff and throughflow occurred, the soil moisture dynamics in the various layers showed a stable period (soil water content is near constant as time elapses) that lasted about 1 week. Also, the pattern of moisture dynamics is ‘decreasing–stabilization–decreasing’. Thus, irrigation and fertilization management according to the spatial and temporal features of soil moisture dynamics on sloped land can increase the water and fertilizer utilization efficacy by reducing their losses during the stable period. Copyright © 2006 John Wiley & Sons, Ltd.     

Streamflow regimes of the Yanhe River under climate and land use change,Loess Plateau,China

Soil and water conservation measures including terracing, afforestation, construction of sediment‐trapping dams, and the ‘Grain for Green Program’ have been extensively implemented in the Yanhe River watershed, of the Loess Plateau, China, over the last six decades, and have resulted in large‐scale land use and land cover changes. This study examined the trends and shifts in streamflow regime over the period of 1953–2010 and relates them to changes in land use and soil and water conservation and to the climatic factors of precipitation and air temperature. The non‐parametric Mann–Kendall test and the Pettitt test were used to identify trends and shifts in streamflow and base flow. A method based on precipitation and potential evaporation was used to evaluate the impacts of climate variability and changes in non‐climate factors changes on annual streamflow. A significant decrease (p = 0.01) in annual streamflow was observed related to a significant change point in 1996, mostly because of significant decreases in streamflow (p = 0.01) in the July to September periods in subsequent years. The annual base flow showed no significant trend from 1953 to 2010 and no change point year, mostly because there were no significant seasonal trends, except for significant decreases (p = 0.05) in the July to September periods. There was no significant trend for precipitation over the studied time period, and no change point was detected. The air temperature showed a significant increasing trend (p < 0.01), and 1986 (p < 0.01) was the change point year. The climate variability, as measured by precipitation and temperature, and non‐climate factors including land use changes and soil and water conservation were estimated to have contributed almost equally to the reduction in annual streamflow. Soil and water conservation practices, including biological measures (e.g. revegetation, planting trees and grass) and engineering measures (such as fish‐scale pits, horizontal trenches, and sediment‐trapping dams) play an important role in reduction of the conversion of rainfall to run‐off. Copyright © 2014 John Wiley & Sons, Ltd.     

Identification of optimal soil hydraulic functions and parameters for predicting soil moisture

Abstract

The accuracy of six combined methods formed by three commonly-used soil hydraulic functions and two methods to determine soil hydraulic parameters based on a soil hydraulic parameter look-up table and soil pedotransfer functions was examined for simulating soil moisture. A novel data analysis and modelling approach was used that eliminated the effects of evapotranspiration so that specific sources of error among the six combined methods could be identified and quantified. By comparing simulated and observed soil moisture at six sites of the USDA Soil Climate Analysis Network, we identified the optimal soil hydraulic functions and parameters for predicting soil moisture. Through sensitivity tests, we also showed that adjusting only the soil saturated hydraulic conductivity, K_s , is insufficient for representing important effects of macropores on soil hydraulic conductivity. Our analysis illustrates that, in general, soil hydraulic conductivity is less sensitive to K_s than to the soil pore-size distribution parameter.

Editor D. Koutsoyiannis; Associate editor D. Hughes

Citation Pan, F., McKane, R.B. and Stieglitz, M., 2012. Identification of optimal soil hydraulic functions and parameters for predicting soil moisture. Hydrological Sciences Journal, 57 (4), 723–737.     

A soil moisture monitoring network to assess controls on runoff generation during atmospheric river events

Soil moisture is a key modifier of runoff generation from rainfall excess, including during extreme precipitation events associated with Atmospheric Rivers (ARs). This paper presents a new, publicly available dataset from a soil moisture monitoring network in Northern California's Russian River Basin, designed to assess soil moisture controls on runoff generation under AR conditions. The observations consist of 2-min volumetric soil moisture at 19 sites and 6 depths (5, 10, 15, 20, 50, and 100 cm), starting in summer 2017. The goals of this monitoring network are to aid the development of research applications and situational awareness tools for Forecast-Informed Reservoir Operations at Lake Mendocino. We present short analyses of these data to demonstrate their capability to characterize soil moisture responses to precipitation across sites and depths, including time series analysis, correlation analysis, and identification of soil saturation thresholds that induce runoff. Our results show strong inter-site Pearson's correlations (>0.8) at the seasonal timescale. Correlations are strong (>0.8) during events with high antecedent soil moisture and during drydown periods, and weak (<0.5) otherwise. High event runoff ratios are observed when antecedent soil moisture thresholds are exceeded, and when antecedent runoff is high. Although local heterogeneity in soil moisture can limit the utility of point source data in some hydrologic model applications, our analyses indicate three ways in which soil moisture data are valuable for model design: (1) sensors installed at 6 depths per location enable us to identify the soil depth below which evapotranspiration and saturation dynamics change, and therefore choose model soil layer depths, (2) time series analysis indicates the role of soil moisture processes in controlling runoff ratio during precipitation, which hydrologic models should replicate, and (3) spatial correlation analysis of the soil moisture fluctuations helps identify when and where distributed hydrologic modelling may be beneficial.     

Potential of soil moisture observations in flood modelling: Estimating initial conditions and correcting rainfall

Rainfall runoff (RR) models are fundamental tools for reducing flood hazards. Although several studies have highlighted the potential of soil moisture (SM) observations to improve flood modelling, much research has still to be done for fully exploiting the evident connection between SM and runoff. As a way of example, improving the quality of forcing data, i.e. rainfall observations, may have a great benefit in flood simulation. Such data are the main hydrological forcing of classical RR models but may suffer from poor quality and record interruption issues. This study explores the potential of using SM observations to improve rainfall observations and set a reliable initial wetness condition of the catchment for improving the capability in flood modelling. In particular, a RR model, which incorporates SM for its initialization, and an algorithm for rainfall estimation from SM observations are coupled using a simple integration method. The study carried out at the Valescure experimental catchment (France) demonstrates the high information content retained by SM for RR transformation, thus giving new possibilities for improving hydrological applications. Results show that an appropriate configuration of the two models allows obtaining improvement in flood simulation up to 15% in mean and 34% in median Nash Sutcliffe performances as well as a reduction of the median error in volume and on peak discharge of about 30% and 15%, respectively.     

对长江宜昌站月平均径流量突变的分析

本文简要叙述了一阶矩突变扫描式t检验和二阶矩突变扫描式F检验算法,并应用于长江宜昌水文站1882-2002年间历年月平均径流量的实测序列,划分出24个年际至十年际时间尺度的相对5级旱、涝时段及其变率特征,与有关长江上游洪水、干旱灾害的历史记载相吻合.     

Streamflow trends and climate linkages in the source region of the Yellow River,China

Much of the discussion on hydrological trends and variability in the source region of the Yellow River centres on the mean values of the mainstream flows. Changes in hydrological extremes in the mainstream as well as in the tributary flows are largely unexplored. Although decreasing water availability has been noted, the nature of those changes is less explored. This article investigates trends and variability in the hydrological regimes (both mean values and extreme events) and their links with the local climate in the source region of the Yellow River over the last 50 years (1959–2008). This large catchment is relatively undisturbed by anthropogenic influences such as abstraction and impoundments, enabling the characterization of widely natural, climate‐driven trends. A total of 27 hydrological variables were used as indicators for the analysis. Streamflow records from six major headwater catchments and climatic data from seven stations were studied. The trend results vary considerably from one river basin to another, and become more accentuated with longer time period. Overall, the source region of the Yellow River is characterized by an overall tendency towards decreasing water availability. Noteworthy are strong decreasing trends in the winter (dry season) monthly flows of January to March and September as well as in annual mean flow, annual 1‐, 3‐, 7‐, 30‐ and 90‐day maxima and minima flows for Maqu and Tangnag catchments over the period 1959–2008. The hydrological variables studied are closely related to precipitation in the wet season (June, July, August and September), indicating that the widespread decrease in wet season precipitation is expected to be associated with significant decrease in streamflow. To conclude, decreasing precipitation, particularly in the wet season, along with increasing temperature can be associated with pronounced decrease in water resources, posing a significant challenge to downstream water uses. Copyright © 2011 John Wiley & Sons, Ltd.