Timothy R. Green

Advances and challenges in predicting agricultural management effects on soil hydraulic properties

Agricultural management practices can significantly affect soil hydraulic properties and processes in space and time. These responses are coupled with the processes of infiltration, runoff, erosion, chemical movement, and crop growth. It is essential to quantify and predict management effects on soil properties in order to model their consequent effects on production and the environment. We present work done thus far on this topic area along with the challenges that lie ahead. The effects of tillage and reconsolidation, wheel-track soil compaction, crop residue management, macropore development and management interactions with natural sources of variability, such as topography, are addressed. Whether explicitly or implicitly, the available field studies include interactions between treatments, such as tillage, crop rotation and residue management. Controlled equipment traffic has been shown to have significant effects on soil compaction and related hydraulic properties in some soils and climates, but in others, landscape and temporal variability overwhelm any effects of wheel tracks. New research results on wheel-track effects in Colorado are highlighted along with initial attempts to predict their effects on hydraulic properties. The greatest challenge for the future is improved process-based prediction using a systems approach to include tightly coupled process interactions in space and time.

A software engineering perspective on environmental modeling framework design: The Object Modeling System

The environmental modeling community has historically been concerned with the proliferation of models and the effort associated with collective model development tasks (e.g., code generation, data transformation, etc.). Environmental modeling frameworks (EMFs) have been developed to address this problem, but much work remains before EMFs are adopted as mainstream modeling tools. Environmental model development requires both scientific understanding of environmental phenomena and software developer proficiency. EMFs support the modeling process through streamlining model code development, allowing seamless access to data, and supporting data analysis and visualization. EMFs also support aggregation of model components into functional units, component interaction and communication, temporal-spatial stepping, scaling of spatial data, multi-threading/multi-processor support, and cross-language interoperability. Some EMFs additionally focus on high-performance computing and are tailored for particular modeling domains such as ecosystem, socio-economic, or climate change research. The Object Modeling System Version 3 (OMS3) EMF employs new advances in software framework design to better support the environmental model development process. This paper discusses key EMF design goals/constraints and addresses software engineering aspects that have made OMS3 framework development efficacious and its application practical, as demonstrated by leveraging software engineering efforts outside of the modeling community and lessons learned from over a decade of EMF development. Software engineering approaches employed in OMS3 are highlighted including a non-invasive lightweight framework design supporting component-based model development, use of implicit parallelism in system design, use of domain specific language design patterns, and cloud-based support for computational scalability. The key advancements in EMF design presented herein may be applicable and beneficial for other EMF developers seeking to better support environmental model development through improved framework design.

The Tarrawarra project: high resolution spatial measurement, modelling and analysis of soil moisture and hydrological response

Detailed spatial patterns of soil moisture were measured for 13 dates at the 10.5 ha Tarrawarra catchment in southern Victoria, Australia. Several analyses of the data are summarized. These include: hydrological behaviour, including preferred states, spatial organization and the performance of terrain indices; geostatistical properties of the soil moisture patterns; and remote sensing of the soil moisture patterns. In the second part of the paper, the patterns along with surface runoff and meteorological data are used in applications of the Thales and VIC models at Tarrawarra. Thales is a process-based distributed parameter hydrological model which explicitly simulates the spatio-temporal patterns of soil moisture, while VIC uses a lumped statistical distribution approach to model the spatial variability of soil moisture storage. Both models simulate saturation excess runoff and are forced by rainfall and potential evapotranspiration. VIC was calibrated to observed runoff at the catchment outlet. Limited manual calibration of Thales to runoff and the soil moisture patterns was performed. Internal testing was achieved by comparison of predicted and observed spatial soil moisture patterns for the Thales model and of predicted and observed cumulative distributions of active soil moisture storage for the VIC model. With limited calibration effort, Thales was able to to simulate the seasonal changes in characteristics of the spatial soil moisture patterns. Detailed examination of the errors in the simulated patterns allowed identification of structural problems in the model, including problems with simulating lateral redistribution as the catchment wets in autumn. For the VIC model, time-series of spatially averaged internal state variables (total storage) were consistent with observations. However, the statistical distribution of soil moisture storage assumed in the model differed from that observed. The collection of detailed spatial data for soil moisture patterns provided a basis for testing the internal states relevant to each model formulation (spatially distributed for Thales and statistically lumped for VIC), as well as improving the identification of the dominant runoff processes.

Relating stream-bank erosion to in-stream transport of suspended sediment

We seek an improved and quantitative understanding of the sources and transport of sediment and attached phosphorus in upland catchments and downstream reaches of the Namoi River in New South Wales, Australia. Study of the sources of phosphorus and related sediment was motivated by severe problems with blooms of blue-green algae and toxic by-products in the Darling and Namoi Rivers. Using atmospheric fall-out of radionuclides as tracers, Olley et al. (1996) concluded that much of the sediment deposited in the lower reaches came from subsoil rather than topsoil. With this insight, we focus on quantifying sediment sources from stream bank erosion, especially in seasonally erosional reaches of Coxs Creek and the Mooki River. The approach presented here integrates interdecadal aerial photography, interseasonal field measurements of bank erosion processes, continuous monitoring of stream flow and turbidity and event sampling of suspended solids and phosphorus, with an analytical model of in-stream suspended sediment transport. We compare a lateral source term in the calibrated transport model with field-based and aerial measurements of stream bank erosion. Calibration of the in-stream model is illustrated for two reaches of the Mooki River, with the changes in parameter values being related to aspects of the hydraulic geometry and particle size. The processes of stream flow and bank erosion due to undercutting, desiccation, block failure and mass wasting of aggregated particles interact to produce instream fluxes of suspended sediment that are transported and redeposited downstream. The combined approach demonstrated here has potential for predictive spatial modelling of sediment concentrations and loads. Copyright

Modelling upland and instream erosion, sediment and phosphorus transport in a large catchment

This overview presents background information to place the subsequent papers by Beavis et al., Dietrich et al. and Green et al. in the context of a unified approach. The modelling framework described here consists of two major components: an upland catchment model and an instream sediment transport model. The upland model simulates stream flow (Q), suspended sediment (SS) and associated phosphorus (P) using rainfall data, and is calibrated to daily stream flow time-series under historical conditions. The instream model routes SS and attached P from the outlet of upland catchments to gauging points downstream. The instream transport model can infer sources (resuspension and bank erosion) and sinks (deposition) within a reach. Aerial photographs are used to assess the on-site effects of climate and land cover/use on erosion and the drainage network. Changes in land cover/use and the effects on the drainage network are related to the parameters in the rainfall-runoff model so that associated effects on Q (and hence SS and P) can be assessed. This modelling framework is prototyped on the Namoi Basin in northern New South Wales, Australia, and is described briefly herein.

State‐Dependent Anisotropy: Comparisons of Quasi‐Analytical Solutions with Stochastic Results for Steady Gravity Drainage

Anisotropy in large-scale unsaturated hydraulic conductivity of layered soils changes with the moisture state. Here, state-dependent anisotropy is computed under conditions of large-scale gravity drainage. Soils represented by Gardners exponential function are perfectly stratified, periodic, and inclined. Analytical integration of Darcy’s law across each layer results in a system of nonlinear equations that is solved iteratively for capillary suction at layer interfaces and for the Darcy flux normal to layering. Computed fluxes and suction profiles are used to determine both upscaled hydraulic conductivity in the principal directions and the corresponding “state-dependent” anisotropy ratio as functions of the mean suction. Three groups of layered soils are analyzed and compared with independent predictions from the stochastic results of Yeh et al. (1985b). The small-perturbation approach predicts appropriate behaviors for anisotropy under nonarid conditions. However, the stochastic results are limited to moderate values of mean suction; this limitation is linked to a Taylor series approximation in terms of a group of statistical and geometric parameters. Two alternative forms of the Taylor series provide upper and lower bounds for the state-dependent anisotropy of relatively dry soils.

Environmental modeling framework invasiveness: Analysis and implications

Environmental modeling frameworks support scientific model development by providing model developers with domain specific software libraries which are used to aid model implementation. This paper presents an investigation on the framework invasiveness of environmental modeling frameworks. Invasiveness, similar to object-oriented coupling, is defined as the quantity of dependencies between model code and a modeling framework. We investigated relationships between invasiveness and the quality of modeling code, and also the utility of using a lightweight framework design approach in an environmental modeling framework. Five metrics to measure framework invasiveness were proposed and applied to measure dependencies between model and framework code of several implementations of Thornthwaite and the Precipitation-Runoff Modeling System (PRMS), two well-known hydrological models. Framework invasiveness measures were compared with existing common software metrics including size (lines of code), cyclomatic complexity, and object-oriented coupling. Models with lower framework invasiveness tended to be smaller, less complex, and have less coupling. In addition, the lightweight framework implementations of the Thornthwaite and PRMS models were less invasive than the traditional framework model implementations. Our results show that model implementations with higher degrees of framework invasiveness also had structural characteristics which previously have been shown to predict poor maintainability, a non-functional code quality attribute of concern. We conclude that using a framework with a lightweight framework design shows promise in helping to improve the quality of model code and that the lightweight framework design approach merits further attention by environmental modeling framework developers.

Soil Moisture Sensing via Swept Frequency Based Microwave Sensors

There is a need for low-cost, high-accuracy measurement of water content in various materials. This study assesses the performance of a new microwave swept frequency domain instrument (SFI) that has promise to provide a low-cost, high-accuracy alternative to the traditional and more expensive time domain reflectometry (TDR). The technique obtains permittivity measurements of soils in the frequency domain utilizing a through transmission configuration, transmissometry, which provides a frequency domain transmissometry measurement (FDT). The measurement is comparable to time domain transmissometry (TDT) with the added advantage of also being able to separately quantify the real and imaginary portions of the complex permittivity so that the measured bulk permittivity is more accurate that the measurement TDR provides where the apparent permittivity is impacted by the signal loss, which can be significant in heavier soils. The experimental SFI was compared with a high-end 12 GHz TDR/TDT system across a range of soils at varying soil water contents and densities. As propagation delay is the fundamental measurement of interest to the well-established TDR or TDT technique; the first set of tests utilized precision propagation delay lines to test the accuracy of the SFI instrument’s ability to resolve propagation delays across the expected range of delays that a soil probe would present when subjected to the expected range of soil types and soil moisture typical to an agronomic cropping system. The results of the precision-delay line testing suggests the instrument is capable of predicting propagation delays with a RMSE of +/−105 ps across the range of delays ranging from 0 to 12,000 ps with a coefficient of determination of r2 = 0.998. The second phase of tests noted the rich history of TDR for prediction of soil moisture and leveraged this history by utilizing TDT measured with a high-end Hewlett Packard TDR/TDT instrument to directly benchmark the SFI instrument over a range of soil types, at varying levels of moisture. This testing protocol was developed to provide the best possible comparison between SFI to TDT than would otherwise be possible by using soil moisture as the bench mark, due to variations in soil density between soil water content levels which are known to impact the calibration between TDR’s estimate of soil water content from the measured propagation delay which is converted to an apparent permittivity measurement. This experimental decision, to compare propagation delay of TDT to FDT, effectively removes the errors due to variations in packing density from the evaluation and provides a direct comparison between the SFI instrument and the time domain technique of TDT. The tests utilized three soils (a sand, an Acuff loam and an Olton clay-loam) that were packed to varying bulk densities and prepared to provide a range of water contents and electrical conductivities by which to compare the performance of the SFI technology to TDT measurements of propagation delay. For each sample tested, the SFI instrument and the TDT both performed the measurements on the exact same probe, thereby both instruments were measuring the exact same soil/soil-probe response to ensure the most accurate means to compare the SFI instrument to a high-end TDT instrument. Test results provided an estimated instrumental accuracy for the SFI of +/−0.98% of full scale, RMSE basis, for the precision delay lines and +/−1.32% when the SFI was evaluated on loam and clay loam soils, in comparison to TDT as the bench-mark. Results from both experiments provide evidence that the low-cost SFI approach is a viable alternative to conventional TDR/TDT for high accuracy applications.

Comparison of electrical and thermal conductivities for soils from five states.

The arrangement of soil particles, particle size, mineralogy, solute concentration, and bulk density affects electrical (&sgr;) and thermal (&lgr;) conductivities, which are key properties for estimating soil physical states, subsurface water and energy balances, and land-atmosphere interactions. The purpose of this study was to compare how &sgr; and &lgr; change as a function of water content for soils under different vegetation and with different properties. Soil samples were collected from selected field sites in Idaho, Texas, Colorado, Iowa, and Ohio and packed into cylinders at a density of 1.2 Mg m−3 and then wetted to predetermined water contents (&thgr;) between 0.10 and 0.45 m3 m−3. A thermo-time domain reflectometer was used to determine &sgr; and &lgr; at each &thgr; at room temperature. Soil and vegetation-influenced differences within a state were only occasionally statistically significant; however, differences between states were highly significant for both &sgr; and &lgr;. The &lgr; decreased as the amount of sorbed water (related to soil-specific surface area) increased. The &lgr; increased more rapidly at low water contents than did &sgr;, but &sgr; increased more rapidly at high water contents. Changes in &sgr; and &lgr; with water content are related to the changing tortuosity of the solid plus liquid phases versus only the liquid phase (including sorbed water). This study contributes toward improved understanding of soil thermal and electrical conductivities over a range of soils.

The Drought Calculator: Decision Support Tool for Predicting Forage Growth During Drought

Abstract The Drought Calculator (DC), a spreadsheet-based decision support tool, was developed to help ranchers and range managers predict reductions in forage production due to drought. Forage growth potential (FGP), the fraction of historical average production, is predicted as a weighted average of monthly precipitation from January through June. We calibrated and evaluated the DC tool in the Great Plains of the United States, using FGP and precipitation data from Colorado (CO), North Dakota (ND), and Wyoming (WY). In CO, FGP was most sensitive to precipitation in April and May, in ND to precipitation in April and June, and in WY to precipitation in April, May, and June. Weights in these months ranged from 0.16 to 0.52. Prediction was better for CO and WY than for ND. When January–June precipitation was used, the tool correctly predicted 83% of the years with FGP reduced by drought for CO, 82% for WY, and only 67% for ND. Positive values of the True Skill Statistic (0.53 for CO, 0.42 for WY, and 0.17 for ND) indicate that FGP was classified as above or below average better than random selection. Predicting FGP earlier than April in CO and WY will require accurate forecasts of April–June precipitation. Use of the DC is most limited by insufficient forage data to determine the site-specific relationships between FGP and monthly precipitation. Even so, the decision tool is useful for discriminating drought effects on FGP classification being above or below the long-term average, and it provides a quantitative prediction to producers for their destocking decisions in drought years.