Soil Water Content Time Series Research Articles

Multistep optimization of HyPix model for flexible vertical scaling of soil hydraulic parameters

Efficient simulation of water-flow processes in the vadose zone is crucial to increase agricultural productivity within environmental limits. This requires deriving detailed soil hydraulic parameters of the soil profile that is highly challenging, particularly for heterogeneous soils. We therefore developed an alternative indirect methodology to calibrate the hydraulic parameters from soil water content time series measured at multiple depths by using the new physically based hydrological model HyPix.We propose a novel, efficient, multistep optimization algorithm for layered soils that derives an optimal set of hydraulic parameters for a desired number of soil layers. For each selected soil layer, HyPix derives five physical, bimodal, Kosugi hydraulic parameters that describe the soil water retention and hydraulic conductivity by using a novel algorithm that reduces the degree of sensitivity and freedom of the parameters. The optimization algorithm upscales the soil hydraulic parameters by gradually incorporating the soil heterogeneity. This method overcomes the problems associated with optimization of the hydraulic parameters of each layer individually, which leads to poor results because it does not represent the cohesive soil water dynamics across the unsaturated zone.We tested the method using soil water content measurements at different depths at five heterogeneous experimental sites in New Zealand. We show how the accuracy of the simulated water balance components increases with the number of soil layers. The multistep optimization upscales a detailed, layered profile of soil hydraulic parameters into a model with fewer layers. The methodology developed provides an estimate of the uncertainty of using a reduced number of soil layers. We also show that a pedological description can provide an indication of the minimum soil layers of vertical discretization required to accurately compute the soil water balance components.

Hierarchical prediction of soil water content time series

Quantifying the spatial and temporal dynamics of soil moisture is an important subject in vadose zone hydrology. A 1282-day field study was conducted to provide a hierarchy of data to assess neural network simulations of field soil water content time series θ(t). Volumetric water content was determined, typically once a week, by neutron thermalization at 20-, 40-, 60-, 80- and 100-cm depths. Soil samples were taken at 60 locations between 14- and 114-cm depths to determine soil properties, water retention, and saturated hydraulic conductivity. Prediction of hydraulic parameters from basic and extended soil properties yielded low correlation coefficients. Water content could be predicted reasonably with neural networks from soil properties or hydraulic parameters (0.880 < R < 0.942). Prediction of θ(t) based solely on rainfall data was not accurate. Independent networks could accurately simulate water content from one observed θ(t) based on the Nash-Sutcliffe Model Efficiency and Percent Bias. Once a network is trained for a particular depth, accurate predictions can be made beyond the training period using observations at just one location.

Long‐range correlations of soil water content time series under tillage and different cover crops in a semi‐arid vineyard

AbstractWe explored the suitability of multifractal detrended fluctuation analysis (MFDFA) to characterize the temporal dynamics of soil water content (SWC), we analysed the origin of the multifractality of time series of the dynamic of SWC and we investigated how this dynamic was affected by different soil treatments and depths, through the characterization of long‐range correlations. Temporal data of SWC from conventional tillage and two different cover crops at three different depths in a Mediterranean vineyard were examined. The dataset spanned 18 months in 30‐min intervals. We investigated if SWC dynamics, as reflected by long‐range dependencies, can be detected and quantified by MFDFA in the form of scaling laws. We analysed the randomly shuffled series associated with temporal data of SWC to investigate if the detected multifractality of the data reflected either a broad density function of the values of the series or different long‐range correlations for small and large fluctuations in the dataset. Finally, the impact of treatment and depth on the variation of long‐range dependencies was studied. Our results suggest that the time records of soil moisture in a Mediterranean vineyard show a multifractal behaviour compatible with complex patterns of long‐range correlations. Moreover, this complex structure is affected by depth, as depicted by the statistically significant variation of the generalized Hurst exponent h(0), which controls the logarithmic mean of the fluctuation of SWC over a period of time. No significant interactions between treatments and depth was found. This suggests that soil properties that influence soil water dynamics at each depth are not affected by cover crop treatments. This could be due to the fact that the experiment lasted less than 2 years and no long‐term interactions were present at this time interval.Highlights Multifractal detrended fluctuation analysis (MFDFA) characterizes soil water content dynamics. MFDFA detects the influence of soil depth on long‐range correlations. Time series of soil moisture are multifractals compatible with complex patterns of long‐range correlations. Long‐range dependencies are affected by depth but no interaction has been found between soil treatment and depth.

Practical Data Products From Cosmic-Ray Neutron Sensing for Hydrological Applications

The Cosmic-Ray Neutron Sensor (CRNS) technique for estimating landscape average soil water content (SWC) is now a decade old and includes now many practical methods for implementing measurements, such as identification of detection area and depth, installation, calibration, and validation. However, in order to maximize the societal relevance of CRNS SWC data, practical value-added products need to be developed that can estimate both water flux (i.e. rainfall, deep percolation, evapotranspiration) and root zone storage changes. In particular, simple methods that can be used to estimate daily values at landscape average scales are needed by decision makers and stakeholders interested in utilizing this technique. Moreover, landscape average values are necessary for better comparisons with remote sensing products. In this work we utilize three well established algorithms to enhance the usability of the CRNS data. The algorithms aim to: 1) temporally smooth the neutron intensity and SWC time series, 2) estimate a daily rainfall product using the Soil Moisture 2 Rain (SM2RAIN) algorithm, and 3) estimate daily root zone SWC using an exponential filter algorithm. The algorithms are tested on the CRNS site at the Hydrological Open Air Laboratory experiment in Petzenkirchen, Austria. Independent observations of rainfall and point SWC data are used to calibrate and validate the algorithms. With respect to rainfall, the SM2RAIN algorithm resulted in a Kling-Gupta-Efficiency (KGE) criteria of 0.665 for daily and 0.819 for 5 day totals. With respect to SWC, the exponential filter algorithm resulted in a KGE of 0.909 for the 0-30cm depth and 0.912 for the 0-60 cm depth. A methodological framework is presented that summarizes the different processes, required data, algorithms, and products.

Disentangling the formation and evolvement mechanism of plants-induced dried soil layers on China’s Loess Plateau

The extent of the effect of China’s Loess Plateau on global climate change partially depends on the achievements of the Grain for Green Program (GFGP). This region is prone to drought problems, especially the widespread creation of dried soil layer (DSL), which is intensified by the revegetation from GFGP initiatives. A DSL is the result of soil desiccation in deep soil profiles. The occurrence and expansion of DSLs are strengthened by the high risk of the Loess Plateau to climate change. On the basis of the 14-year data of soil water content (SWC) of four revegetation types on a loessial slope, we investigated the evolution processes and recovery possibility of DSL. Results showed that both variation trends and abrupt changes in the SWC time series were distinctly affected by revegetation. The 40 cm thickness of the DSL was triggered when Medicago sativa and Caragana korshinskii grew for 2 and 4 years, respectively. The downward extension depth of DSL reached 580 cm under the two plants above within 7 and 8 years of planting. However, the results suggested the possibility of DSL disappearance as shown by the formation depth dynamics of DSL, the fluctuation of DSL-SWC, and the phenomenon of sandwiched DSL. During deep moisture recharge and DSL recovery, although the SWC sensitivity to textural variation was present in the vertical profile, the effects of soil texture on SWC were decreased. Wet climate was linked to both DSL thinning and thickening. Thus, the effect of wet climate is bidirectional for DSL development, thereby highlighting the important role of revegetation. The results of this work enhance our understanding of DSL and may help alleviate the risk of drought in the Loess Plateau.

Temporal dynamics of subsurface soil water content estimated from surface measurement using wavelet transform

Temporal variation of water in the vadose zone is important to understand processes such as solute transport and nutrient cycling. Measurements of soil water content (SWC) in the subsurface are less common than those near the surface and predictions using numerical models are limited by data availability. Wavelet decomposition of surface measurements of SWC could improve modeling of subsurface SWC by segregating features at different temporal scales and projecting them to the subsurface. The objectives of this work were to: 1) predict subsurface SWC using surface SWC and a combination of wavelet analysis and linear regression, and 2) investigate relationships between soil properties and the movement of soil water at various temporal scales, s. Climate data and SWC at various depths were collected from eight sites in the Atlantic Coastal Plain of the USA. Soil water retention and hydraulic conductivity (k) functions of each horizon were optimized by comparing measured and predicted (using HYDRUS-1D) soil water contents. Each time series of SWC was decomposed into 50 scale components using the Mexican Hat wavelet and later reduced to 5 group components with minimal impact on the characteristics of the signal. Changes in the values of each group component with depth were represented with transfer coefficients that could be estimated with predictors derived from particle size distributions and optimized soil hydraulic functions. Prediction depth and saturated k were the two most important predictors for s < 256 h, while k at −10 kPa was the best predictor for 256 h < s < 724 h, and the median value of particle size diameters for s > 724 h. Subsurface soil water content can be reasonably predicted with the proposed approach, particularly when vertical movement of soil water is unrestricted.

Evaluating the Timing and Interdependence of Hydrologic Processes at the Watershed Scale Based on Continuously Monitored Data

A quantitative understanding of the interplay between the different components of the hydrologic cycle at the watershed scale can be gained from analyzing high-frequency hydrologic time series. High-frequency measurements of precipitation, soil water content, shallow groundwater, and streamflow were collected and analyzed in Otter Creek, a 122 km2 watershed located in Northeast Iowa, USA. For selected rainfall events occurring in 2014, it was found that there is at least 4 h of delay between soil water content and water table time series response and streamflow peak. This is true even when the water table was approximately 6.5 m below the ground surface before rainfall started. Data reveal a strong linear dependence between the soil water content and the water table, which suggests the existence of a capillary fringe that extends approximately 2.5 m above the water table. The highest streamflow values in Otter Creek occurred when both the water table was close to the ground surface and the near surface soil (top 65 cm) was close to full saturation. Analyses show that, in the study area, data on depth to water table or deep soil water content have the potential to play a key role in the development of a flood warning system. The transformation of rainfall into streamflow is a complex process that we simplified in this study. Additional analyses using physically based coupled surface-subsurface models or non-linear or stochastic models are recommended for more rigorous analysis.

Estimating Field Capacity from Volumetric Soil Water Content Time Series Using Automated Processing Algorithms

Core Ideas We evaluated approaches based on machine learning and soil water dynamic principles. We compared automated analysis of soil moisture sensor time series for estimating field capacity. Approaches accurately estimated field capacity compared with expert values. Automated approaches needed to leverage temporally dynamic field capacity. Vadose zone measurements of volumetric soil water content (θ) using soil moisture sensors (SMSs) have become more common due to advances in technology and reduction of costs. Soil moisture sensor data exhibit a characteristic cyclical pattern reflecting water flux dynamics into and out of the observed soil volume. Expert review of SMS datasets to distinguish valid from corrupt or incomplete soil water cycles is arguably the most precise method for determining field capacity (θFC) but is impractically cumbersome and time consuming for increasingly large SMS datasets. We evaluated competing approaches for automated soil water cycles analysis that use widely available R packages based on pattern recognition and machine learning (findpeaks [R‐FP], symbolic aggregate approximation [R‐SAX], and density histogram [R‐DH]), and a MATLAB code based on soil water dynamic principles (SWDP). These approaches were applied to three SMS datasets. Our empirical results showed superiority of R‐SAX for identifying valid soil water cycles, probably due to benefiting from training sets to calibrate to correct cycles. Two other approaches (SWDP and R‐FP) provided similar results without need of training sets or preprocessing data. Three approaches for estimating field capacity were applied to valid cycles, R‐FP, regression of exponential decay (SWDP‐R), and estimated “knee” of curve (SWDP‐K). Each performed similarly to the expert defined values, with R‐FP and SWDP‐R generally performing best across analyses. Results of this study also show temporal dynamics of θFC within datasets used here. There is potential for optimizing θFC and a need for automated, objective analysis to leverage dynamics in irrigation management and modeling.

Water Flow Modeling with Dry Bulk Density Optimization to Determine Hydraulic Properties in Mountain Soils

Core Ideas The GEOtop model was used to calculate water flow in mountain soils. Dry soil bulk density was optimized to determine soil hydraulic properties. Soil water contents for two watersheds were used to test the optimization procedure. A new method for determining profile‐average and depth‐wise hydraulic properties in heterogeneous mountain soils is presented using the GEOtop watershed model in 1‐D vertical mode. Dry soil bulk density–converted volumetric soil water retention data are used to determine van Genuchten soil water retention parameters, and the Kozeny–Carman equation is used to determine saturated soil hydraulic conductivity. Optimum dry soil bulk densities are identified by minimizing the sum of squared error between measured and calculated soil water content time series. The new method was tested using soil moisture data from soil profiles at the Dry Creek Experimental Watershed, Boise, ID, and the Libby Creek Experimental Watershed, Laramie, WY. Results of different scenarios showed that the optimization of a single profile‐average dry soil bulk density is a good option for describing soil water flow in the heterogeneous mountain soils. Soil water content modeling efficiency (ME) values of 0.084 ≤ ME ≤ 0.745 and –2.443 ≤ ME ≤ 0.373 were found for the Dry Creek and Libby Creek sites, respectively. Relatively low ME values for the deepest sensor depths for some scenarios were attributed to the overestimation of soil water freezing and uncertainty in the soil water retention function near saturation. The resulting calibration procedure is computationally efficient because only one parameter (dry soil bulk density) is optimized.

Erratum to “Disentangling the Effects of Land Management and Soil Heterogeneity on Soil Moisture Dynamics”

The following three passages and the respective references were not included in the published version. They should have been added to the paragraphs specified below. 4 studied the effects of different land management practices on deep seepage and solute leaching based on soil water content and tension time series measured in 3 m soil depth. The field experiment was designed to study the effects of land use and land management on deep seepage and solute leaching (3; 1). Time series of soil moisture and soil water tension were continually measured in high temporal solution, and soil water samples were taken in different soil depths down to 3 m. The measurements started in 1994 and continue today. Our findings about the small functional heterogeneity at the experimental site confirmed the basic assumption by 2 that relative soil moisture dynamics at different positions in the deep vadose zone are very similar. They developed and tested a field method for quantifying deep seepage and solute leaching based on tension and water content measurements and applied it successfully at 40 plots in northeast Germany (3; 4).

Using measured soil water contents to estimate evapotranspiration and root water uptake profiles – a comparative study

Abstract. Understanding the role of plants in soil water relations, and thus ecosystem functioning, requires information about root water uptake. We evaluated four different complex water balance methods to estimate sink term patterns and evapotranspiration directly from soil moisture measurements. We tested four methods. The first two take the difference between two measurement intervals as evapotranspiration, thus neglecting vertical flow. The third uses regression on the soil water content time series and differences between day and night to account for vertical flow. The fourth accounts for vertical flow using a numerical model and iteratively solves for the sink term. None of these methods requires any a priori information of root distribution parameters or evapotranspiration, which is an advantage compared to common root water uptake models. To test the methods, a synthetic experiment with numerical simulations for a grassland ecosystem was conducted. Additionally, the time series were perturbed to simulate common sensor errors, like those due to measurement precision and inaccurate sensor calibration. We tested each method for a range of measurement frequencies and applied performance criteria to evaluate the suitability of each method. In general, we show that methods accounting for vertical flow predict evapotranspiration and the sink term distribution more accurately than the simpler approaches. Under consideration of possible measurement uncertainties, the method based on regression and differentiating between day and night cycles leads to the best and most robust estimation of sink term patterns. It is thus an alternative to more complex inverse numerical methods. This study demonstrates that highly resolved (temporally and spatially) soil water content measurements may be used to estimate the sink term profiles when the appropriate approach is used.

Upper washita river experimental watersheds: multiyear stability of soil water content profiles.

Scaling in situ soil water content time series data to a large spatial domain is a key element of watershed environmental monitoring and modeling. The primary method of estimating and monitoring large-scale soil water content distributions is via in situ networks. It is critical to establish the stability of in situ networks when deploying them to study hydrologic systems. Two watersheds in Oklahoma, the Little Washita River Experimental Watershed (LWREW) and the Fort Cobb Reservoir Experimental Watershed (FCREW), are two prime examples of well-equipped research watersheds that provide long-term measurements of atmospheric and soil water content from in situ networks. The soil water content measurement network on the LWREW has been in operation since 2002, with 20 stations available for investigating soil water dynamics at the watershed scale. Temporal stability analysis of the network is complicated by the changing configuration of the network, but it is possible to determine a singular long-term average for the network. The FCREW consists of 15 soil water content stations and began operation in 2007, providing detailed information across a mixed agricultural domain and was determined to be stable and representative of the region. This study reinforces the applicability of temporal stability analysis to very long time scales, which are now becoming available for soil moisture monitoring. Each of these networks is temporally stable with respect to soil water content at each depth on a spatial basis. The LWREW has a persistent pattern through the root zone profile, but the FCREW does not, which requires further investigation.

Modelling soil water dynamics considering measurement uncertainty

AbstractIn shallow water table‐controlled environments, surface water management impacts groundwater table levels and soil water dynamics. The study goal was to simulate soil water dynamics in response to canal stage raises considering uncertainty in measured soil water content. Water and Agrochemicals in the soil, crop and Vadose Environment (WAVE) was applied to simulate unsaturated flow above a shallow aquifer. Global sensitivity analysis was performed to identify model input factors with the greatest influence on predicted soil water content. Nash–Sutcliffe increased and Root Mean Square Error reduced when uncertainties in measured data were considered in goodness‐of‐fit calculations using measurement probability distributions and probable asymmetric error boundaries, implying that appropriate model performance evaluation should be carried out using uncertainty ranges instead of single values. Although uncertainty in the experimental measured data limited evaluation of the absolute predictions by the model, WAVE was found a useful exploratory tool for estimating temporal variation in soil water content. Visual analysis of soil water content time series under proposed changes in canal stage management indicated that sites with land surface elevation of less than 2.0‐m NGVD29 were predicted to periodically experience saturated conditions in the root zone and shortening of the growing season if canal stage is raised more than 9 cm and maintained at this level. The models developed could be combined with high‐resolution digital elevation models in future studies to identify areas with the greatest risk of experiencing saturated root zone. The study also highlighted the need to incorporate measurement uncertainty when evaluating performance of unsaturated flow models. Copyright © 2014 John Wiley &amp; Sons, Ltd.

Time series outlier and intervention analysis: Irrigation management influences on soil water content in silty loam soil

Understanding the field soil water regime is fundamental in scheduling irrigation as well as for monitoring water flow and solute transport. This study was carried out on variable interval irrigation and used time series analysis techniques to predict the soil water content at the interested depth by measuring one single depth in order to precisely determine the next irrigation time and its effect on soil water content at the interested depth. Volumetric water content of silty loam soil in Barcelona was measured in situ with capacitance soil moisture sensors at five depths within the root zone for a horticultural crop during its life cycle in 2010. The time series consisted of hourly measurements of soil water content and was transformed to a stationary situation. Subsequently, the transformed data were used to conduct analyses in the time domain in order to obtain the parameters of a seasonal autoregressive integrated moving average (ARIMA) model. In the case of variable interval irrigation, predicting the soil water content time series cannot be properly explained by the ARIMA model and its underlying normality assumption. By completing the ARIMA model with intervention analysis and outlier detection, the prediction of soil water content in variable interval irrigation can be made. The transfer function models were then used to predict water contents at depths of interest (0.20, 0.35, 0.50 and 0.60m depths) as well as the average water content WAVG in the top 0.60m soil profile by measuring water content at 0.10m depth. As a result, the predictions were logical. Also, the next irrigation time and its effect on soil water content at the depth of interest were correctly estimated. To confirm results of the models, the experiment was repeated in 2011, and the predicted and observed values agree reasonably well.

Event-based estimation of water budget components using a network of multi-sensor capacitance probes

A new approach was developed for estimating vertical soil water fluxes using soil water content time series data. Instead of a traditional fixed time interval, this approach utilizes the time interval between two sequential minima of the soil water storage time series to identify groundwater recharge events and calculate components of the soil water budget. We calculated water budget components: surface-water excess (Sw), infiltration less evapotranspiration (I – ET) and groundwater recharge (R) from May 2001 to January 2003 at eight locations at the USDA Agricultural Research Center, Beltsville, Maryland, USA. High uncertainty was observed for all budget components. This uncertainty was attributed to spatial and temporal variation in Sw, I – ET and R, and was caused by nonuniform rainfall distributions during recharge events, variability in the profile water content, and spatial variability in soil hydraulic properties. The proposed event-based approach allows estimating water budget components when profile water content monitoring data are available. Citation Guber, A., Gish, T., Pachepsky, Y., McKee, L., Nicholson, T. & Cady, R. (2011) Event-based estimation of water budget components using a network of multi-sensor capacitance probes. Hydrol. Sci. J. 56(7), 1227–1241.

An empirical model for the evaluation of the degree of saturation of shallow soils in relation to rainfalls

The evaluation of shallow soils water content is very important in many fields, and different hydrological models are widely applied to simulate field-scale water flow and soil water content. The degree of saturation of a shallow soil is a time-varying variable, depending on different weather conditions. In many applications, it is useful to directly correlate the soil degree of saturation to time series of rainfall amount. This paper presents a simplified empirical model, which allows for computation of the soil degree of saturation using readily available climate data on air temperature and rainfall depths. The model is tested with in situ measurements of soil water content collected at three sites in the Emilia Romagna region in northern Italy. The experimental data are compared with the results obtained from the simplified model over an observation period of almost 5 years. The model has been used to generate soil water content time series at different depths and requires a separate calibration for each depth where the degree of saturation is computed. A discussion on the model’s calibration is also carried out to clarify the model sensitivity to the different parameters and the choice of input data.

Artificial neural network and time series models for predicting soil salt and water content

Abstract Volumetric water content of a silt loam soil (fluvo-aquic soil) in North China Plain was measured in situ by L-520 neutron probe (made in China) at three depths in the crop rootzone during a lysimeter experiment from 2001 to 2006. The electrical conductivity of the soil water (ECsw) was measured by salinity sensors buried in the soil during the same period at 10, 20, 45 and 70 cm depth below soil surface. These data were used to test two mathematical procedures to predict water content and soil water salinity at depths of interest: all the available data were divided into training and testing datasets, then back propagation neural networks (BPNNs) were optimized by sensitivity analysis to minimizing the performance error, and then were finally used to predict soil water and ECsw. In order to meet with the prerequisite of autoregressive integrated moving average (ARIMA) model, firstly, original soil water content and ECsw time series were likewise transformed to obtain stationary series. Subsequently, the transformed time series were used to conduct analysis in frequency domain to obtain the parameters of the ARIMA models for the purposes of using the ARIMA model to predict soil water content and ECsw. Based on the statistical parameters used to assess model performance, the BPNN model performed better in predicting the average water content than the ARIMA model: coefficient of determination (R2) = 0.8987, sum of squares error (SSE) = 0.000009, and mean absolute error (MAE) = 0.000967 for BPNN as compared to R2 = 0.8867, SSE = 0.000043, MAE = 0.002211 for ARIMA. The BPNN model also performed better than the ARIMA model in predicting average ECsw of soil profile. However, the ARIMA model performed better than the BPNN models in predicting soil water content at the depth of 20 cm and ECsw at the depth of 10 cm below soil surface. Overall, the model developed by BPNN network showed its advantage of less parameter input, nonlinearity, simple model structure and good prediction of soil ECsw and water content, and it gave an alternative method in forecasting soil water and salt dynamics to those based on deterministic models based on Richards’ equation and Darcy's law provided climatic, cropping patterns, salinity of the irrigation water and irrigation management are very similar from one year to the next.

SOIL WATER FLUCTUATIONS IN A LOAMY SAND UNDER IRRIGATED POTATO

Diagnostic indicators were developed to evaluate water storage, infiltration, drainage, and uptake distribution patterns over a potato (Solanum tuberosum L.) hill to improve efficiency in drip and sprinkler irrigation. Soil water content time series were measured in the hills and in the furrow between hills using time domain reflectometry. The drainage indicator was significantly (P = 0.05) higher under sprinkler than drip beneath the furrow and at the toe location. Storage was low at the center under sprinkler irrigation (0.05 m3 m−3) where irrigation water did not infiltrate. Water storage was low at the shoulder location under both irrigation systems (significantly greater for sprinkler than drip) where high uptake rates were indicated. A diurnal/sinusoidal model of water uptake was fitted to a composite of the daily cycle of low-amplitude fluctuations with acceptable results (R2 = 0.39-0.72) for both systems at the shoulder and toe locations and at the center under drip. Uptake was virtually undetectable (R2 = 0.12-0.18) at the base and furrow locations. Deviations from the sinusoidal model were better accounted for by a polynomial model (R2 = 0.29-0.83) and seemed to indicate irrigation-induced flow and nonsinusoidal uptake. Electronic noise also contributed to the variability. The indicator of peak uptake time shifted from about 3 p.m. to 8 p.m., and uptake decreased as the depth from the top of the hill increased. Management practices targeted at wetting the hill center under the sprinkler would likely improve water use efficiency under these conditions.

State–space prediction of field–scale soil water content time series in a sandy loam

The description of field soil water content time series can be affected by uncertainty due to measurement errors of the respective state variables, errors due to assumptions underlying the model, and errors in the determination of boundary conditions. In this study, a simple state-equation was applied for predicting field soil water contents at three different soil depths. The simple state-model yielded large deviations of predictions from the measured soil water content, especially for the upper soil depth. Apparently, the magnitude of the estimated evaporation rate was too high. The prediction result could significantly be improved when the calculated evaporation was reduced by a factor of 0.7. In order to account for uncertainty sources associated with this simple approach, the state-equation was combined with a stochastic technique, the so-called Kalman–Filter. Applying the Kalman–Filter, the prediction quality significantly increased, even when the erroneously high evaporation was assumed to be true. However, prediction uncertainty increased for the same time periods, for which it was shown earlier that spatial correlation of soil water status was either random or very short. When the Kalman–Filter was applied in a scenario to the surface layer only, simulated soil water content in layers 2 and 3 agreed to measurements and were highly improved compared to simulations when layer 1 was not filtered. Hence, application of lab determined soil hydraulic property functions in combination with state observations of upper soil horizon water content and with the Kalman–Filter provides a promising opportunity to describe and predict soil water contents for entire soil profiles even under the presence of uncertainty sources.

Time Series Analysis of Field‐Measured Water Content of a Sandy Soil

AbstractPrediction of the field soil water regime is important for assessing agrochemical transport and scheduling irrigation. This research was conducted to evaluate the potential for employing the time series analysis technique to predict average water content (WAVG) of a soil profile and water content at depths of interest from measurements made at a single depth. Volumetric water content of a Zimmerman fine sand (mixed, frigid Argic Udipsamment) in Princeton, MN, was measured in situ by time domain reflectometry at six depths during the early 1993 growing season. The time series made up of hourly measurements of soil water content was first‐order differenced to obtain stationarity. The differenced data were used to conduct analyses in the frequency domain to evaluate the coherence and cross‐amplitude between two soil water content time series and were subsequently fitted to autoregressive moving average models to obtain coefficients for the transfer function models in the time domain. The transfer function models were then used to predict water contents at depths of interest (50‐, 75‐, and 100‐cm depths) and WAVG in the top 100‐cm profile from measured water content at the 25‐cm depth. Overall, the predictions were reasonable, with accuracy increasing as the separation distance from the 25‐cm depth decreased.