There are a great number of publications that apply different methods to estimate soil moisture from optical satellite imagery. However, none of the proposed methods have considered correcting solar illumination error that is caused by variation in topography before estimating soil moisture. In this research, an integrated approach is developed to improve the estimation of soil moisture. The integration is represented by removing the solar-illumination error from the data. Several modifications were made in the Ts-VI space based on the Universal Triangle Relationship. The data used in the research are obtained from Moderate Resolution Imaging Spectroradiometer (MODIS) satellite. The research results show that the surface-illumination error, which is caused by variation in topography, misleads the estimation of soil moisture index. Based on statistical and visual analysis, the results are improved with removing error. The method is further enhanced with the application of enhanced vegetation index (EVI) to the Ts-VI relationship.

This study sought to investigate the application of artificial neural networks (ANN) and fuzzy inference systems (FIS) to variably saturated soil moisture (VSSM) redistribution modelling. An enhanced approach to such modelling, that lessens computation costs, facilitates input preparation, handles data uncertainty, and realistically simulates soil moisture redistribution, was our main objective. / An initial review of existing soil hydrology models provided greater insight into current modelling challenges and a general classification of the models. The application of AI techniques as alternative tools for soil hydrology modelling was explored. / A one-dimensional (1D) model based on ANN and FIS was developed. To estimate fluxes more accurately, multiple ANNs were trained and combined by way of an FIS. The main body of the model employed the ANN-FIS module to model soil moisture redistribution throughout the profile. When tested against the SWAP93 model, the ANN-FIS model gave a good match and maximum error of <8%; however, it did not show a notable computation cost shift. / The investigation proceeded with development of another ANN-based 1D modelling approach. This time, the soil profile or flow region, regardless of its depth, was divided into ten equal parts (compartments). The ANN was trained to estimate moisture patterns for a whole soil profile, from the previous day's soil moisture pattern and boundary conditions, and the current day's boundary conditions. The model was tested against SWAP93 where an average SCORE of 90.4 indicated a good match. The computation cost of the ANN-based model was about one-third that of SWAP93. / At this point the study sought to develop a 3D modelling approach. The ANN was trained to estimate the nodal soil moisture changes through time under the influence of six neighbouring nodes (in a 3D space, two on each axis). The model's accuracy was tested against the SWMS-3D model. An average SCORE of 91 and a 15-fold decrease in computation costs showed a quite acceptable performance. Results suggest that this approach is potentially capable of realistically modelling 3D VSSM redistribution with less computation time. / Finally, pros and cons of these ANN-based modelling approaches are compared and contrasted, and some recommendations on future work are given.

No description available.

Three studies were undertaken from 2006 to 2007 to examine litter (dead plant material) in southwestern Manitoba pastures. First, the relationship between litter and soil microclimate was tested across five pasture sites. The amount of litter biomass was not strongly related to soil moisture, though near-surface soil temperatures were reduced when litter was present. Second, the effect of four simulated grazing strategies on the litter layer was measured in six pastures. It was found that after three years of simulated grazing, litter was present in largest quantities in the least-frequently grazed treatments. Finally, a field survey was conducted assessing the quantity of litter present in native pastures across Manitoba. Litter was quite variable and averaged 1902 kg/ha over two years. This research confirmed the value of litter as an indicator of sustainable pasture management, though it remains unclear whether litter is important to pastures from the perspective of soil microclimate.

To assess the accuracy of the second-generation Prairie Agrometeorological Model (PAM2nd) as an agrometeorological model for spring wheat on the Canadian Prairies, a study was conducted to validate the model using field measurements. Results from model validation indicated soil moisture was being overestimated at most sites during the second half of the growing season, while soil moisture was underestimated during periods that experienced consecutive days of rainfall. Modifications to the model were implemented to improve the model's ability to simulate soil moisture. Evapotranspiration estimates from PAM2nd and the FAO56 Penmen-Monteith method were compared to water balance methods. Both models produced estimates that fell within the range of water balance ET measurement error. The similarity in performance of both models to estimate ET compared to the water balance ET means the adoption of either model could be justified. However, PAM2nd would be more appropriate because it requires fewer, more commonly measured, surface weather parameters.

A computer simulation model of the soil water regime can be a useful research, planning and management tool, providing that the data required by the model are available. Finite difference solutions of the general flow equation can be applied to complex field situations if soil profile characteristics are reflected by appropriate retentivity (B( Ψ)) and hydraulic conductivity (K(Ψ)) functions. The validity of a flow simulation model depends upon the degree to which simulated flow corresponds to the flow pattern in real soils. Macroscopic flow in apedal soils is likely to obey Darcy's law but in structured or swe~ling soils, macro-pores and shrinkage voids lead to non-Darcian flow. Physical composition and structural stability properties of a wide range of South African soils were used to assess swelling behaviour and depth-related textural changes. The applicability of a one-dimensional Darcian flow model to various soil types was evaluated. Core retentivity data for South African soils were used to derive regression equations for predicting B (Ψ) from textural criteria and bulk density. A sigmoidal, non-hysteretic two-part retentivity function having only two constants in addition to porosity was developed for use in water flow simulation models. Values of the constants, shapes of the retentivity curves and soil textural properties were related by fitting the retentivity function to retentivity data generated using regression equations~ Hydraulically inhomogeneous soils may be modelled by varying the values of the retentivity constants through the profile to reflect changing soil properties. Equations for calculating K(B) or K(Ψ) from retentivity data were derived by applying each of three capillary models to both exponential and two-part retentivity functions. Comparison of these equations showed that the definition and value of semi-empirical constants in the capillary models were as important as the choice of model in determining K(B). K(Ψ) was calculated using retentivity constants corresponding to a range of bulk density, clay and silt content combinations. Three retentivity constant-soil property systems were evaluated. These were derived from retentivity data for South African soils between 1) -10 and -1500 kPa, 2) 0 and -50 kPa and 3) from published retentivity data for British soils. Only that derived from retentivity data accurate in the 0 to -50 kPa range led to K(Ψ) relationships in which saturated K and the slope sK/sΨ decreased as bulk density, clay or silt content increased. Absolute values of K were unreliable and measured values are essential for matching purposes. A method for evaluating the constants in a K(Ψ) or K(B) function from the rate of outflow or inflow of water after a step change in potential at the base of a soil core was described. Simple exponential g (Ψ) and K(Ψ) functions were assumed to apply to each pressure potential range. Retentivity parameters were obtained by fitting the 8(Ψ) function to the measured retentivity curve. A value for K[s] , the remaining unknown parameter in the K(Ψ) function, was obtained by matching measured outflow and inflow data to a family of simulated curves. These were computed using measured retentivity parameters, core dimensions and ceramic plate conductivity, and a range of K[s] values. An advantage of this method is that there are no limitations on core length, plate impedance or pressure potential range which cannot be ascertained by prior simulation. Regression equations relating texture to retentivity, and a conductivity model were applied in a simulation study of the water regime in a weighing lysimeter in which gains and losses of water were measured accurately. Active root distribution was assumed proportional to root mass distribution. Relative K(Ψ) curves for each node were computed using one of the conductivity equations derived earlier. Daily water potentials for a month were simulated using three conductivity matching factors. By matching simulated Ψ values to tensiometer potentials measured at five depths an appropriate matching factor was chosen. The effects of an over- or underestimate of K(Ψ) were demonstrated. This work simplifies the prediction and use of retentivity and conductivity relationships in soil water flow simulation models. These models can be used for assessing the water regime in both irrigated and dry-land crop production. Other applications include catchment modelling, effluent disposal and nutrient and solute transport in soil. / Thesis (Ph.D.)-University of Natal, Pietermaritzburg, 1983.

Three studies were undertaken from 2006 to 2007 to examine litter (dead plant material) in southwestern Manitoba pastures. First, the relationship between litter and soil microclimate was tested across five pasture sites. The amount of litter biomass was not strongly related to soil moisture, though near-surface soil temperatures were reduced when litter was present. Second, the effect of four simulated grazing strategies on the litter layer was measured in six pastures. It was found that after three years of simulated grazing, litter was present in largest quantities in the least-frequently grazed treatments. Finally, a field survey was conducted assessing the quantity of litter present in native pastures across Manitoba. Litter was quite variable and averaged 1902 kg/ha over two years. This research confirmed the value of litter as an indicator of sustainable pasture management, though it remains unclear whether litter is important to pastures from the perspective of soil microclimate.

To assess the accuracy of the second-generation Prairie Agrometeorological Model (PAM2nd) as an agrometeorological model for spring wheat on the Canadian Prairies, a study was conducted to validate the model using field measurements. Results from model validation indicated soil moisture was being overestimated at most sites during the second half of the growing season, while soil moisture was underestimated during periods that experienced consecutive days of rainfall. Modifications to the model were implemented to improve the model's ability to simulate soil moisture. Evapotranspiration estimates from PAM2nd and the FAO56 Penmen-Monteith method were compared to water balance methods. Both models produced estimates that fell within the range of water balance ET measurement error. The similarity in performance of both models to estimate ET compared to the water balance ET means the adoption of either model could be justified. However, PAM2nd would be more appropriate because it requires fewer, more commonly measured, surface weather parameters.

Note: / ln May 1962 the McGill University Savanna Research Project wasestablished and has been conducted since that date in the Department ofGeography p McGill University and in the savannas of the Rupununi Di strict pBritish Guiana and the Territorio do Rio Brancop Brazi!.It is generally recognised that although a very wide range of theoryhas been propounded to explain the nature and origin of savannas p no onehas yet brought forward a single convincing viewpoint which has met withuniversal acceptance. One reason for this is that the majority of theoriesextant are based upon inadequate fie ld data with almost a total lack ofexperimental evidence. The McGill Univers ity Sa vanna Research Projectwas set up for the purpose of initiating an experimental and observati onalfield programme which it was hoped would shed light upon the ecologicalrelations of the savanna p and would ultimately enable an explanation to bemade of the nature and distribution of the savanna vegetation of the region .

Effective management of saline soils requires rapid, reliable methods of monitoring both soil water content (theta) and salt concentration, the latter measured in terms of electrical conductivity (sigma). This thesis examines estimation of theta, and bulk soil sigma (sigmaa) and soil water sigma (sigmaw), using time domain reflectometry (TDR). Calibration experiments were conducted in soil columns and in a vineyard that were irrigated with saline water. Within a theta range of 0.025--0.490 m3/m3, the correlation between TDR theta (thetaTDR) and gravimetrically determined theta (theta g) was high (r2 = 0.979 in soil columns, r2 = 0.836 in the field). The error of estimate of thetaTDR was 0.020 m3/m3 or less. Field thetaTDR estimates were sensitive to high salinity (sigmaw > 10 dS/m). Using a dual pathway parallel conductance (DPPC) model, sigma a was derived from sigmaw of saturated paste extracts and theta g. The correlation of TDR sigmaa to DPPC sigmaa was good in the laboratory (r2 = 0.915), and moderate in the field (r2 = 0.791), indicating a functional relationship between sigmaw and theta and TDR sigmaa. Models, to estimate sigmaw, were built by regression between paste extract sigma w and TDR sigmaa and thetaTDR. In a sigma w range of 3.0--23.4 dS/m in the columns, error of estimate of sigma w was small at 1.50 dS/m (+/-12.4% relative error range). In a sigma w range of 2.2--25.2 dS/m in the field, error of estimate of sigma w was 3.37 dS/m (+/-37% relative error range), which was significantly higher than the acceptable error range of +/-10%. ANOVA tests indicated that both TDR sigmaa and sigmaw&middot;theta changed significantly with respect to the same sources of variance. Error in field estimates of sigmaw was associated with effects of salinity on thetaTDR and variability of soil conditions, particularly with respect to depth and time of sampling.