Tesis sobre el tema "Soil moisture"

Soil moisture information can be used accurately in determining the timing and amount of irrigation applied to plants. Pan and Pan et al. proposed a robust and simple daily diagnostic equation for estimating daily soil moisture. The diagnostic equation evaluates the relationship between the soil moisture loss function and the summation weighted average of precipitation. The loss function uses the sinusoidal wave function which employs day of the year (DOY) to evaluate the seasonal variation in soil moisture loss for a given year. This was incorporated into the daily diagnostic equation to estimate the daily soil moisture for a location. Solar radiation is an energy source that drives the energy and water exchanges between vegetation and the atmosphere (i.e., evapotranspiration), and thus impacts the soil moisture dry-down. In this paper, two parameters (the actual solar radiation and the clear sky solar radiation) are introduced into loss function coefficient to improve the estimation of soil moisture. After the Introduction of the solar radiation data into soil moisture loss function, a slight improvement was observed in the estimated daily soil moisture. Pan observed that generally the correlation coefficient between the estimated and the observed soil moisture is above 0.75 and the root mean square error is below 5.0 (%v/v). The introduction solar radiation data (i.e. clear sky solar radiation and actual solar) improve the correlation coefficient average for all the sites evaluated by 0.03 when the root mean square error is generally below 4.5(%v/v) for the entire root zone.

Thesis (M.S.)--George Mason University, 2008.
Vita: p. 65. Thesis director: John Qu. Submitted in partial fulfillment of the requirements for the degree of Master of Science in Earth System Science. Title from PDF t.p. (viewed Jan. 11, 2009). Includes bibliographical references (p. 59-64). Also issued in print.

The reliability and accuracy of several sensors that employ the relationship between dielectric constant and soil moisture constant, e, in particular capacitance sensors were investigated. Results obtained from laboratory examinations ,of a Theta probe, TP, selected as a representative model for capacitance sensors, suggested that the sensor output was affected by temperature variations, electrical conductivity levels, spatial variation in sample bulk density as well as the level of compaction of the soil surrounding the sensor's rods. Detailed in situ e data collected usmg capacitance sensors were used to calculate sub-daily estimates of evaporation, E, using the soil water balance method, combined with the zero-flux-plane (ZFP) approach, for plots of bare soil, rapeseed and a maize field. These sensors comprised Theta probes (TP), Profiles probes (PP), ECH20 probes (EP) and Aquaflex sensors (AF). / The field output data of these sensors were analysed and compared with e obtained with both, the gravimetric and neutron probe method. The absolute values of B as measured by the various capacitance sensors differed considerably. Furthermore, the outputs of these sensors (apart from the AF probes) were found to be affected by temperature, which would result in an anomalous course of diurnal E. Also, B-data were subject to noise which required smoothing to ensure a physically realistic variation in E, when compared to estimates with the Penman-Monteith equation, EPAf, and the eddy-covariance method (maize field). E was determined from diurnal changes in vertically integrated soil moisture content above the ZFP. Smoothed values of Bwere temperature-corrected using fieldbased and laboratory-based correction equations. A considerable difference between field- and laboratory-based temperature corrections procedures was noticed, and correction factors strongly depended on B. As this resulted in an overly complicated correction procedure, which consequently gave unreliable E-values, it was then decided to use a constant correction factor (based on the field correction procedure), for each capacitance probe. For the bare soil plot, with the exception ofPP and EP only Bprofiles obtained with the TP and AF sensors produced relatively reliable E values when compared to Enf. By contrast, when these capacitance sensors were used under a canopy, all sensors yielded satisfactory E-values. This was most likely caused by reduced amplitudes of soil temperatures under the canopy and the fact that the dimensions of most sensors do not allow installation in the top soil (~3-5cm) layer at which most evaporation would take place in bare soils. We therefore recommended that these sensors can be used for diurnal B measurements and E determination under canopy provided that an appropriate temperature-correction procedure for each sensor is applied. To obtain reliable Band E estimates in bare soil, more research needs to be done. For more reliable e and E estimations in bare soils further extensive field trials would be strongly advised

The use of soil cement as a growth medium was examined in this study. During the monitoring, green soil cement revealed diverse ecological values. The survival rates of plants in each soil conditions were higher than 80%,which was very promising. Furthermore, the survival rates dropped when the soil density reached95%, which means soil density might influence the survival rate of plant. Plant growth rates in sandy soil were higher than that in soil cement. In particular, low soil density facilitated plant growth in sandy soil, whereas density effect was not clear to plant growth performance in soil cement. Experiments were undertaken to study the soil-water characteristics of sandy soil and soil cement in field and laboratory condition. The influence of vegetation and material density on the development of negative pore water pressure (PWP) and degree of saturation (Sr) in the studied materials was investigated. The field planting experiments proved a promising survival rate of Schefflera heptaphylla in both types of materials while sandy soil promoted better growth of the seedlings than the soil cement. From the field study, PWP and Sr of sandy soil responded noticeably and promptly to natural drying and wetting cycles. However, the responses in soil cement were relatively mild. When subjected to the same drying-wetting cycles, PWP responded more slowly and to a smaller magnitude compared with that of soil cement. In addition, Sr changed little in soil cement. An increase in the density of the sandy soil promoted rapid development of negative PWP, while an opposite trend was observed for soil cement. Attempts have been made to explain the observations from the perspectives of material permeability and change in water content during a drying period in both soil types. Furthermore, in sandy soil, the development of PWP (with a measurement limit of -90 kPa) was minimally affected by the presence of vegetation, while vegetation noticeably helped the development of negative PWP in the soil cement. Bounds of the soil-water characteristic curve of the studied materials were presented based on estimates from the drying and wetting scanning curves derived from the field monitoring. A complementary laboratory study was carried out in an environmental chamber with controllable temperature and humidity. Monitoring results from the laboratory agreed well with that obtained from the field.
published_or_final_version
Civil Engineering
Master
Master of Philosophy

Synthetic Aperture Radar (SAR) has been identified as a good candidate to provide high-resolution soil moisture information over extended areas. SAR data could be used as observations within a global Data Assimilation (DA) approach to benefit applications such as hydrology and agriculture. Prior to developing an operational DA system, one must tackle the following challenges of soil moisture estimation with SAR: (1) the dependency of the measured radar signal on both soil moisture and soil surface roughness which leads to an ill-conditioned inverse problem, and (2) the difficulty in characterizing spatially/temporally surface roughness of natural soils and its scattering contribution. The objectives of this project are (1) to develop a roughness measurement method to improve the spatial/temporal characterization of soil surface roughness, and (2) to investigate to what extent the inverse problem can be solved by combining multipolarization, multi-incidence, and/or multi-frequency radar measurements. The first objective is achieved with a measurement method based on Structure from Motion (SfM). It is tailored to monitor natural surface roughness changes which have often been assumed negligible although without evidence. The measurement method is flexible, a.ordable, straightforward and generates Digital Elevation Models (DEMs) for a SAR-pixel-size plot with mm accuracy. A new processing method based on band-filtering of the DEM and its 2D Power Spectral Density (PSD) is proposed to compute the classical roughness parameters. Time series of DEMs show that non-negligible changes in surface roughness can happen within two months at scales relevant for microwave scattering. The second objective is achieved using maximum likelihood fitting of the Oh backscattering model to (1) full-polarimetric Radarsat-2 data and (2) simulated multi-polarization / multi-incidence / multi-frequency radar data. Model fitting with the Radarsat-2 images leads to poor soil moisture retrieval which is related to inaccuracy of the Oh model. Model fitting with the simulated data quantifies the amount of multilooking for di.erent combinations of measurements needed to mitigate the critical e.ect of speckle on soil moisture uncertainty. Results also suggest that dual-polarization measurements at L- and C-bands are a promising combination to achieve the observation requirements of soil moisture. In conclusion, the SfM method along with the recommended processing techniques are good candidates to improve the characterization of surface roughness. A combination of multi-polarization and multi-frequency radar measurements appears to be a robust basis for a future Data Assimilation system for global soil moisture monitoring.

If given the correct remotely sensed information, machine learning can accurately describe soil moisture conditions in a heterogeneous region at the large scale based on soil moisture readings at the small scale through rule transference across scale. This paper reviews an approach to increase soil moisture resolution over a sample region over Australia using the Soil Moisture Active Passive (SMAP) sensor and Landsat 8 only and a validation experiment using Sentinal-2 and the Advanced Microwave Scanning Radiometer (AMSR-E) over Nevada. This approach uses an inductive localized approach, replacing the need to obtain a deterministic model in favor of a learning model. This model is adaptable to heterogeneous conditions within a single scene unlike traditional polynomial fitting models and has fixed variables unlike most leaning models. For the purposes of this analysis, the SMAP 36 km soil moisture product is considered fully valid and accurate. Landsat bands coinciding in collection date with a SMAP capture are down sampled to match the resolution of the SMAP product. A series of indices describing the Soil-Vegetation-Atmosphere Triangle (SVAT) relationship are then produced, including two novel variables, using the down sampled Landsat bands. These indices are then related to the local coincident SMAP values to identify a series of rules or trees to identify the local rules defining the relationship between soil moisture and the indices. The defined rules are then applied to the Landsat image in the native Landsat resolution to determine local soil moisture. Ground truth comparison is done via a series of grids using point soil moisture samples and air-borne L-band Multibeam Radiometer (PLMR) observations done under the SMAPEx-5 campaign (Panciera 2013). This paper uses a random forest due to its highly accurate learning against local ground truth data yet easily understandable rules. The predictive power of the inferred learning soil moisture algorithm did well with a mean absolute error of 0.054 over an airborne L-band retrieved surface over the same region. The validation experiment also demonstrated a strong linkage to the soil moisture, but the algorithm suffered from a lack of training data over such a small site. However, soil moisture estimation still exhibited a mean average error (MAE) of 0.028, compared to a 0.129 MAE of a deterministic model built upon the Air Force Weather Model.

One key component in good irrigation management is the measurement of soil moisture to help determine when to irrigate. In this study, resistance blocks and tensiometers were compared to neutron probe readings to assess how well these devices followed soil moisture and whether the resistance blocks and /or tensiometers could be used to schedule irrigation in cotton production. The resistance blocks were placed at 6, 18, and 30 inches. Tensiometers were placed at 18 and 30 inches. The readings from the resistance blocks and tensiometers were compared to neutron probe readings taken at 6, 18, and 30 inches. The resistance blocks compared well with the neutron probe readings at the 6 inch and 30 inch depth. At the 18 inch depth, there was much scatter in the data. The tensiometers also showed good comparisons at 30 inches and poor comparisons at 18 inches.

Soil moisture is a key state variable that controls runoff formation, infiltration and partitioning of radiation into latent and sensible heat. However, the experimental characterisation of near surface soil moisture patterns and their controls on runoff formation remains a challenge. This subject was one aspect of the BMBF-funded OPAQUE project (operational discharge and flooding predictions in head catchments). As part of that project the focus of this dissertation is on: (1) testing the methodology and feasibility of the Spatial TDR technology in producing soil moisture profiles along TDR probes, including an inversion technique of the recorded signal in heterogeneous field soils, (2) the analysis of spatial variability and temporal dynamics of soil moisture at the field scale including field experiments and hydrological modelling, (3) the application of models of different complexity for understanding soil moisture dynamics and its importance for runoff generation as well as for improving the prediction of runoff volumes. To fulfil objective 1, several laboratory experiments were conducted to understand the influence of probe rod geometry and heterogeneities in the sampling volume under different wetness conditions. This includes a detailed analysis on how these error sources affect retrieval of soil moisture profiles in soils. Concerning objective 2 a sampling strategy of two TDR clusters installed in the head water of the Wilde Weißeritz catchment (Eastern Ore Mountains, Germany) was used to investigate how well “the catchment state” can be characterised by means of distributed soil moisture data observed at the field scale. A grassland site and a forested site both located on gentle slopes were instrumented with two Spatial TDR clusters that consist of up to 39 TDR probes. Process understanding was gained by modelling the interaction of evapotranspiration and soil moisture with the hydrological process model CATFLOW. A field scale irrigation experiment was carried out to investigate near subsurface processes at the hillslope scale. The interactions of soil moisture and runoff formation were analysed using discharge data from three nested catchments: the Becherbach with a size of 2 km², the Rehefeld catchment (17 km²) and the superordinate Ammelsdorf catchment (49 km²). Statistical analyses including observations of pre-event runoff, soil moisture and different rainfall characteristics were employed to predict stream flow volume. On the different scales a strong correlation between the average soil moisture and the runoff coefficients of rainfall-runoff events could be found, which almost explains equivalent variability as the pre-event runoff. Furthermore, there was a strong correlation between surface soil moisture and subsurface wetness with a hysteretic behaviour between runoff soil moisture. To fulfil objective 3 these findings were used in a generalised linear model (GLM) analysis which combines state variables describing the catchments antecedent wetness and variables describing the meteorological forcing in order to predict event runoff coefficients. GLM results were compared to simulations with the catchment model WaSiM ETH. Hereby were the model results of the GLMs always better than the simulations with WaSiM ETH. The GLM analysis indicated that the proposed sampling strategy of clustering TDR probes in typical functional units is a promising technique to explore soil moisture controls on runoff generation and can be an important link between the scales. Long term monitoring of such sites could yield valuable information for flood warning and forecasting by identifying critical soil moisture conditions for the former and providing a better representation of the initial moisture conditions for the latter.
Abflussentwicklung, Infiltration und die Umverteilung von Strahlung in latenten und sensiblen Wärmestrom werden maßgeblich durch die Bodenfeuchte der vadosen Zone gesteuert. Trotz allem, gibt s wenig Arbeiten die sich mit der experimentellen Charakterisierung der Bodenfeuchteverteilung und ihre Auswirkung auf die Abflussbildung beschäftigen. Der Fokus dieser Dissertation wurde darauf ausgerichtet: (1) die Methode des Spatial TDR und deren Anwendbarkeit einschließlich der Inversion des TDR Signals in heterogenen Böden zu prüfen, (2) die Analyse der räumlichen und zeitlichen Dynamik der Bodenfeuchte auf der Feldskala einschließlich Feldexperimenten und hydrologischer Modellierung, (3) der Aufbau verschiedener Modellanwendungen unterschiedlicher Komplexität um die Bodenfeuchtedynamiken und die Abflussentwicklung zu verstehen und die Vorhersage des Abflussvolumens zu verbessern. Um die Zielsetzung 1 zu erreichen, wurden verschiedene Laborversuche durchgeführt. Hierbei wurde der Einfluss der Sondenstabgeometrie und verschiedener Heterogenitäten im Messvolumen bei verschiedenen Feuchtegehalten untersucht. Dies beinhaltete eine detaillierte Analyse wie diese Fehlerquellen die Inversion des Bodenfeuchteprofils beeinflussen. Betreffend der Zielsetzung 2, wurden 2 TDR-Cluster in den Quellgebieten der Wilden Weißeritz installiert (Osterzgebirge) und untersucht, wie gut der Gebietszustand mit räumlich hochaufgelösten Bodenfeuchtedaten der Feldskala charakterisiert werden kann. Um die Interaktion zwischen Evapotranspiration und Bodenfeuchte zu untersuchen wurde das hydrologische Prozessmodell CATFLOW angewendet. Ein Beregnungsversuch wurde durchgeführt um die Zwischenabflussprozesse auf der Hangskala zu verstehen. Die Interaktion zwischen Bodenfeuchte und Abflussentwicklung wurde anhand von drei einander zugeordneten Einzugsgebieten analysiert. Statistische Analysen unter Berücksichtigung von Basisabfluss, Bodenvorfeuchte und verschiedenen Niederschlagscharakteristika wurden verwendet, um auf das Abflussvolumen zu schließen. Auf den verschiedenen Skalen konnte eine hohe Korrelation zwischen der mittleren Bodenfeuchte und dem Abflussbeiwert der Einzelereignisse festgestellt werden. Hierbei konnte die Bodenfeuchte genauso viel Variabilität erklären wie der Basisabfluss. Im Hinblick auf Zielsetzung 3 wurden “Generalised liner models” (GLM) genutzt. Dabei wurden Prädiktorvariablen die den Gebietszustand beschreiben und solche die die Meteorologische Randbedingungen beschreiben genutzt um den Abflussbeiwert zu schätzen. Die Ergebnisse der GLMs wurden mit Simulationsergebnissen des hydrologischen Gebietsmodells WaSiM ETH verglichen. Hierbei haben die GLMs eindeutig bessere Ergebnisse geliefert gegenüber den WaSiM Simulationen. Die GLM Analysen haben aufgezeigt, dass die verwendete Messstrategie mehrerer TDR-Cluster in typischen funktionalen Einheiten eine viel versprechende Methode ist, um den Einfluss der Bodenfeuchte auf die Abflussentwicklung zu verstehen und ein Bindeglied zwischen den Skalen darstellen zu können. Langzeitbeobachtungen solcher Standorte sind in der Lage wichtige Zusatzinformationen bei der Hochwasserwarnung und -vorhersage zu liefern durch die Identifizierung kritischer Gebietszustände für erstere und eine bessere Repräsentation der Vorfeuchte für letztere.

Land surface albedo is the ratio of reflected to incident solar radiation. It is a function of several surface parameters including soil color, moisture, roughness and vegetation cover. A better understanding of albedo and how it changes in relation to variations in these parameters is important in order to help improve our ability to model the effects of land surface modifications on climate. The objectives of this study were 1) To determine empirical relationships between smooth bare soil albedo and soil color, 2) To develop statistical relationships between albedo and ground-based thematic mapper (TM) measurements of spectral reflectances, 3) To determine how increased surface roughness caused by tillage reduces bare soil albedo and 4) To empirically relate albedo with TM data and other physical characteristics of mixed grass/shrubland sites at Walnut Gulch Watershed. Albedos, colors and spectral reflectances were measured by Eppley pyranometer, Chroma Meter CR-200 and a Spectron SE-590, respectively. Measurements were made on two field soils (Gila and Pima) at the Campus Agricultural Center (CAC), Tucson, AZ. Soil surface roughness was measured by a profile meter developed by the USDA/ARS. Additional measurements were made at the Maricopa Agricultural Center (MAC) for statistical model testing. Albedos of the 15 smooth, bare soils (plus silica sand) were determined by linear regression to be highly correlated (r²=0.93, p>0.01) with color values for both wet and dry soil conditions. Albedos of the same smooth bare soils were also highly correlated (r²≥0.86, p>0.01) with spectral reflectances. Testing of the linear regression equations relating albedo to soil color and spectral reflectances using the data from MAC showed a high correlation. A general nonlinear relationship given by y=8.3661n(x)+37.802 r²=0.71 was determined between percent reduction in albedo (y) and surface roughness index (x) for wet and dry Pima and Gila field soils. Measurements of albedo, color and spectral reflectance at the Walnut Gulch Watershed indicated that albedo values were highly correlated with percent rock & gravel, color value and reflectance data (TM bands 1-4).

Thesis (Ph. D.)--Hong Kong University of Science and Technology, 2003.
Includes bibliographical references (leaves 471-490). Also available in electronic version. Access restricted to campus users.

Thesis (Ph. D.) -- University of Maryland, College Park, 2005.
Thesis research directed by: Meteorology. Title from t.p. of PDF. Includes bibliographical references. Published by UMI Dissertation Services, Ann Arbor, Mich. Also available in paper.

This study investigated the possibility of three approaches in determination of soil moisture in the root zone. The aim of the study was to contribute to the development of soil moisture monitoring methods to better help crop best management practices.
Two fields were examined, one at the Macdonald Campus of McGill University and the other near St. Jean-sur-Richelieau. Three approaches were used; (1) a hand-held hyper-spectral sensor (350-2500 nm), (2) a Geonics RTM EM-38 conductivity meter and, (3) gravimetric soil moisture sampling.
The first experiment (at St. Jean-sur-Richelieu) investigated the possibility of monitoring soil moisture with the EM_38, in the presence of field elevation and soil texture. The second experiment (at Macdonald Campus) investigated the possibility of using hyper-spectral sensor data for determination of soil characteristics in the root zone, in the presence of such factors as (a) irrigation (main treatment), (b) nitrogen (sub-treatment), and (c) weed control (sub-sub-treatment). Statistical regression analyses and Artificial Neural Network models were used to select the best waveband region for determination of soil root zone moisture.
The coefficients of determination obtained by the statistical analyses ranged from 0.75 to 0.94. The wavebands most frequently identified by these analyses ranged from 1100 nm-1900 nm.
The performances of the ANN training models were considered acceptable (R2 from 0.6 to 0.8). The lack of sufficient data greatly impacts this approach.

In this thesis an automated greenhouse was built with the purpose of investigating the watering system’s reliability and if a desired range of temperatures can be maintained. The microcontroller used to create the automated greenhouse was an Arduino UNO. This project utilizes two different sensors, a soil moisture sensor and a temperature sensor. The sensors are controlling the two actuators which are a heating fan and a pump. The heating fan is used to change the temperature and the pump is used to water the plant. The watering system and the temperature control system was tested both separately and together. The result showed that the temperature could be maintained in the desired range. Results from the soil moisture sensor were uneven and therefore interpret as unreliable.
I denna tes byggdes ett automatiserat växthus med syftet att undersöka dess bevattningssystems pålitlighet samt om ett önskat temperaturspann kan bibehållas. Microkontrollern för att bygga detta automatiserade växthus var en Arduino UNO. Detta projekt använder sig av två olika sensorer, en jordfuktsensor och en temperatursensor. Sensorerna kontrollerar en värmefläkt och en pump. Värmefläkten används för att ändra temperaturen och pumpen för att vattna plantan. Bevattningssystemet och temperaturstyrningen har testats både separat och tillsammans. Resultatet visar att temperaturen kan bibehållas inom det önskade spannet. Resultaten från jordfuktsensorn var ojämna och därför tolkats som opålitliga.

An IoT device measuring soil moisture to help a municipality improve the work surrounding irrigation is deployed for testing and used by the concerned municipal workers. The IoT device is an initial prototype using Pycom’s LoPy4 with expansion board 3.1 and 3xAAA batteries as power supply. The prototype is not suitable for larger-scale testing due to the size, cost, and power consumption. This thesis focuses on decreasing the cost and size while increasing the battery life for the IoT device.The IoT device is communicating using the LoRaWAN protocol. For the device to be as energy-efficient as possible, the LoRa and LoRaWAN protocol are explored to use all possibilities to save energy. Active time and power consumption between different spreading factors are examined and discussed for power consumption and range concerns. The prototype produced by the work performed in this thesis is used to measure soil moisture. The microcontrollers selected and tested will not restrict to that purpose. The microcontrollers can integrate with many kinds of sensors. Integration with other types of sensors is under future work in this thesis. The integration possibilities making the thesis relevant for anyone with an intermediate knowledge inprogramming wanting to get introduced into the IoT development cycle, develop aLoRa node, and learning how to use the LoRaWAN stack with MicroPython.

Satellite radiometers are used to remotely measure properties of the Earth's surface. Radiometers enable wide spatial coverage and daily temporal coverage. Radiometer measurements are used in a wide array of applications, including freeze/thaw states inference, vegetation index calculations, rainfall estimation, and soil moisture estimation. Resolution enhancement of these radiometer measurements enable finer details to be resolved and improve our understanding of Earth. The Soil Moisture Active Passive (SMAP) radiometer was launched in April 2014 with a goal to produce high resolution soil moisture estimates. However, due to hardware failure of the radar channels, prepared algorithms could no longer be used. Current algorithms utilize a narrow spatial and temporal overlap between the SMAP radiometer and the SENTINEL-1 radar to produce high resolution soil moisture estimates that are spatially and temporally limited. This thesis explores the use of resolution enhancing algorithms to produce high resolution soil moisture estimates without the spatial coverage limitations caused by using multiple sensors. Two main approaches are considered: calculating the iterative update in brightness temperature and calculating the update in soil moisture. The best performing algorithm is the Soil Moisture Image Reconstruction (SMIR) algorithm that is a variation of the Radiometer form of the Scatterometer Image Reconstruction (rSIR) algorithm that has been adapted to operate in parameter space. This algorithm utilizes a novel soil moisture measurement response function (SMRF) in the reconstruction. It matches or exceeds the performance of other algorithms and allows for wide spatial coverage.

Thesis (Ph.D.)--George Mason University, 2008.
Vita: p. 151. Thesis director: John J. Qu. Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Earth Systems and GeoInformation Sciences. Title from PDF t.p. (viewed June 30, 2008). Includes bibliographical references (p. 135-150). Also issued in print.

Thesis (MScAgric)--University of Stellenbosch, 2004.
ENGLISH ABSTRACT: Irrigation scheduling is one of the most important cultural practices in irrigated vineyards. Water holding capacity of soil is arguably therefore one of the most important characteristics of a soil as it determines how much water can be made available to the plant. The measurement of water holding capacity of soils is time consuming and costly. In situ determinations are often impractical to determine. For routine determinations, water holding capacity is therefore determined on disturbed samples. Such a method for example is the rubber ring method. A great deal of criticism surrounds this rubber ring method and results are often questioned. The objectives of this study were therefore to determine what the relationship was between undisturbed and disturbed samples and to determine whether compacted samples could give a more accurate representation of the water holding capacity of soil. Soil textural factors influencing the volumetric water content of undisturbed, rubber ring and compacted samples at 5, 10 and 100 kPa were investigated. In addition, soil textural properties influencing water holding capacity of the respective samples between 5 and 100 kPa and 10 and 100 kPa were investigated. The final objective of the study was to develop simple models to predict the volumetric water content and water holding capacity of soil. Undisturbed and disturbed soil samples were taken at various localities to ensure a wide range of textures. Water holding capacity of undisturbed and disturbed samples was determined at ARC Infruitec-Nietvoorbij using the standard air pressure and ceramic plate technique and the routine rubber ring method respectively. Soil samples were also compacted to a bulk density of approximately 1.5 g.cm-3 as a further treatment for determination of water holding capacity using the air pressure and ceramic plate technique. To investigate aspects of soil texture that could possibly influence volumetric water content of the soil, correlations were done between different texture components and volumetric water content of undisturbed, rubber ring and compacted samples at 5, 10 and 100 kPa. In order to determine the effect of texture on water holding capacity of the soil, correlations were drawn between texture components and water holding capacity of undisturbed, rubber ring and compacted samples between matric potential ranges 5 and 100 kPa and 10 and 100 kPa. The results from this study were used to develop models to predict volumetric soil water content and water holding capacity of soils for a range of soils. Volumetric water content of rubber ring samples at 5 kPa was more than the volumetric water content of undisturbed samples at 5 kPa. The volumetric water content of rubber ring samples at 5 kPa and the volumetric water content of undisturbed samples at 5 kPa was correlated by 87%. Volumetric water content of compacted samples at 5 kPa had a 85% degree of correlation with the volumetric water content of undisturbed samples. At 10 kPa, the correlation between volumetric water content determined using rubber ring samples and undisturbed samples was 77%. This was identical to the correlation between volumetric water content of compacted samples at 10 kPa and undisturbed samples. At 100 kPa, most of the rubber ring samples' volumetric water content fell below the 1:1 line of volumetric water content of undisturbed samples. The volumetric water content of all the compacted samples was higher than that of the undisturbed samples. Water holding capacity of all the rubber ring samples between 5 and 100 kPa was greater than the water holding capacity of the undisturbed samples between 5 and 100 kPa. Rubber ring samples therefore generally overestimated the water holding capacity of the soil. The water holding capacity of most of the rubber ring samples between 10 and 100 kPa was greater than the water holding capacity of the undisturbed samples. In contrast, the water holding capacity of compacted samples between 5 and 100 kPa was less than the water holding capacity of undisturbed samples between 5 and 100 kPa. Water holding capacity of compacted samples was therefore underestimated. The results from this study confirmed that the influence of clay and silt content on volumetric water content of undisturbed, rubber ring and compacted samples increased as the suction on the respective samples is increased. The influence of fine sand content on volumetric water content of undisturbed, rubber ring and compacted samples decreased with an increase in matric potential to 100 kPa. Medium sand content of undisturbed, rubber ring and compacted samples had the greatest influence of all the textural components on the volumetric water content of the respective samples at 5 kPa and 10 kPa. Water holding capacity of undisturbed, rubber ring and compacted samples between 5 and 100 kPa was greatly influenced by the fine sand content of the samples. Medium sand content of the samples also had an influence on the water holding capacity thereof. To predict the volumetric water content of undisturbed samples at 5, 10 and 100 kPa, the independent variables were fine sand content, square root of medium sand content and In of medium sand content. In the case of models to predict the volumetric water content of rubber ring samples at 5, 10 and 100 kPa, the same variables were used as independent variables. Additional variables such as silt content, the In of silt content, square root of clay plus silt content and the medium sand content. To predict the volumetric water content of compacted samples at 5, 10 and 100 kPa the terms used were silt content, clay plus silt content, the e-clay plus silt content. medium sand content and the square root of medium sand content. The models to predict volumetric water content of rubber ring samples gave the best correlation with the actual volumetric water content of rubber ring samples. The final models to predict the water holding capacity of all the samples between 5 and 100 kPa and 10 and 100 kPa used only fine and medium sand parameters as independent variables. Soil textural components do play an important role in determining the volumetric water content of undisturbed, rubber ring and compacted samples at 5, 10 and 100 kPa. The magnitude of the water holding capacity between 5 and 100 kPa and 10 and 100 kPa is also influenced by soil texture. The models developed to predict the volumetric water content of samples at 5, 10 and 100 kPa and the magnitude of the water holding capacity between 5 and 100 kPa and 10 and 100 kPa could be very useful. Both time and money can potentially be saved. Models that can be highly recommended are the models generated for the undisturbed samples. These are: At 5 kPa, VWCu = 0.47259 - 0.04712 medium sando.s At 10 kPa, VWCu = 0.41292 - 0.04221 medium sandos At 100 kPa, VWCu = 0.48080 - 0.00254 fine sand - 0.0865 In medium sand Between 5 and 100 kPa, WHCu = -29.523 + 3.394 fine sand Between 10 and 100 kPa, WHCu = -891.794 + 232.326 In fine sand + 38.006 In medium sand
AFRIKAANSE OPSOMMING: Besproeiingskedulering is een van die belangrikste wingerdverbouingspraktyke. Waterhouvermoë bepaal hoeveel water beskikbaar gestel kan word aan die plant en daarom is dit een van die belangrikste eienskappe van 'n grond. Die meting van waterhouvermoë van grond is tydsaam en duur. Boonop is in situ bepalings dikwels onprakties om te bepaal. Waterhouvermoë word dus bepaal op versteurde monsters vir roetine ontledings. 'n Voorbeeld van so 'n metode is die rubberring metode. Daar bestaan groot kritiek teenoor hierdie rubberring metode en resultate word dikwels betwyfel deur die landboubedryf. Die doel van hierdie studie was dus om te bepaal wat die verwantskap is tussen onversteurde monsters en rubberring monsters asook om te bepaal of gekompakteerde monsters 'n meer akkurate aanduiding sou gee as onversteurde monsters van die waterhouvermoë van die grond. Grondtekstuur faktore wat die volumetriese waterinhoud van onversteurde monsters, rubberring monsters en gekompakteerde monsters by 5, 10 and 100 kPa beïnvloed, was ondersoek. Grondtekstuur faktore wat waterhouvermoë van die onderskeie monsters tussen 5 en 100 kPa en tussen 10 en 100 kPa beïnvloed, was ook ondersoek. Die finale doelwit van die studie was om eenvoudige modelle te ontwikkel vir die voorspelling van volumetriese waterinhoud en waterhouvermoë van grond. Onversteurde grond monsters en grond vir versteurde monsters is by verskeie lokaliteite geneem om 'n wye reeks teksture te verkry. Waterhouvermoë van onversteurde monsters is bepaal by LNR Infruitec- Nietvoorbij met die standaard drukplaat tegniek. Waterhouvermoë van versteurde grond is bepaal met die roetine rubberring metode van LNR Infruitec-Nietvoorbij. Grond was ook gekompakteer tot 'n bulkdigtheid van ongeveer 1.5 g.cm-3 en daarna is die waterhouvermoë bepaal by die LNR Infruitec- Nietvoorbij met die standaard drukplaat tegniek. Om aspekte van grondtekstuur, wat moontlik die volumetriese waterinhoud van grond kan beïnvloed te ondersoek, is korrelasies tussen verskeie tekstuur komponente en die volumetriese waterinhoud van onversteurde monsters, rubberring monsters en gekompakteerde monsters by 5, 10 en 100 kPa bepaal. Om te bepaal watter tekstuur komponente waterhouvermoë van die grond kan bepaal, is korrelasies getrek tussen tekstuur komponente en waterhouvermoë van onversteurde monsters, rubberring monsters en gekompakteerde monsters tussen 5 en 100 kPa en tussen 10 en 100 kPa. Die data is verwerk met die SAS uitgawe 6.12 (SAS, 1990) om modelle vir die voorspelling van volumetriese waterinhoud en waterhouvermoë van grond met behulp van maklik kwantifiseerbare grondtekstuur veranderlikes te ontwikkel. Die volumetriese waterinhoud van rubberring monsters by 5 kPa was meer as die volumetriese waterinhoud van onversteurde monsters by 5 kPa. Die volumetriese waterinhoud van rubberring monsters by 5 kPa en die volumetriese waterinhoud van onversteurde monsters by 5 kPa is gekorreleerd met 87%. Die volumetriese waterinhoud van gekompakteerde monsters by 5 kPa het 'n korrelasie van 85% met volumetriese waterinhoud van onversteurde monsters getoon. By 10 kPa, was die graad van korrelasie tussen volumetriese waterinhoud bepaal met rubberring monsters en onversteurde monsters, 77%. Dit was omtrent dieselfde as die graad van korrelasie tussen volumetriese waterinhoud van gekompakteerde monsters en onversteurde monsters by 10 kPa. By 100 kPa het die meeste van die rubberring monsters se volumetriese waterinhoud onderkant die 1:1 lyn van die volumetriese waterinhoud by 100 kPa van al die onversteurde monsters. Die volumetriese waterinhoud van al die gekompakteerde monsters was hoër as die van die onversteurde monsters. Die waterhouvermoë van al die rubberring monsters tussen 5 en 100 kPa was groter as die van die onversteurde monsters tussen 5 en 100 kPa. Die rubberring monsters het dus oor die algemeen die grootte van die waterhouvermoë oorskry. Die waterhouvermoë van die meeste van die rubberring monsters tussen 10 en 100 kPa was groter as die waterhouvermoë van die onversteurde monsters. Die waterhouvermoë van gekompakteerde monsters tussen 5 en 100 kPa was minder as die waterhouvermoë van die onversteurde monsters tussen 5 en 100 kPa. Die waterhouvermoë van gekompakteerde grondmonsters is dus onderskat. Die resultate van hierdie studie het die invloed van klei- en slik- inhoud op die volumetriese waterinhoud van onversteurde monsters, rubberring monsters en gekompakteerde monsters bevestig. Die invloed van klei en sand op die volumetriese waterinhoud van onversteurde monsters, rubberring monsters en gekompakteerde monsters het toegeneem soos die matriks potensiaal op die onderskeie monsters toegeneem het. Die invloed van fynsand op die volumetriese waterinhoud van onversteurde monsters, rubberring monsters en gekompakteerde monsters was die grootste by 5 kPa en het afgeneem tot by 100 kPa. Die mediumsand inhoud van onversteurde monsters, rubberring monsters en gekompakteerde monsters het van al die tekstuur komponente die grootste invloed op die volumetriese waterinhoud van al die monsters by 5 kPa en 10 kPa gehad. Die waterhouvermoë van onversteurde monsters, rubberring monsters en gekompakteerde monsters tussen 5 en 100 kPa is grootliks beinvloed deur die fynsand inhoud van die monsters. Die mediumsand inhoud van die monsters het ook 'n invloed gehad op die waterhouvermoë daarvan. Om die volumetriese waterinhoud van onversteurde monsters by 5, 10 en 100 kPa te voorspel, is onafhanklike veranderlikes soos fynsand inhoud, vierkantswortel van mediumsand inhoud en In van mediumsand inhoud bepaal. In die geval van modelle om die volumetriese waterinhoud van rubberring monsters by 5, 10 en 100 kPa te voorspel, is dieselfde veranderlikes gebruik as onafhanklike veranderlikes. Addisionele veranderlikes soos slik inhoud, In van slik inhoud, die vierkantswortel van die klei plus slik inhoud en die mediumsand inhoud is ook gebruik. Om die volumetriese waterinhoud van gekompakteerde monsters by 5, 10 en 100 kPa te voorspel, is die terme slik inhoud, klei plus slik inhoud, e-klei plus slik inhoud, mediumsand inhoud en vierkantswortel van mediumsand inhoud gebruik. Die modelle om volumetriese waterinhoud van rubberring samples te voorspel het die akkuraatste voorspellings gegee. Die finale modelle, om waterhouvermoë van alle monsters tussen 5 en 100 kPa en tussen 10 en 100 kPa te bepaal, het slegs fyn en mediumsand as onafhanklike veranderlikes gebruik. Grondtekstuur komponente speel dus 'n belangrike rol in die volumetriese waterinhoud van onversteurde monsters, rubberring monsters en gekompakteerde monsters by 5, 10 en 100 kPa. Die grootte van die waterhouvermoë tussen 5 en 100 kPa en tussen 10 en 100 kPa is ook beinvloed deur die grondtekstuur. Die modelle wat ontwikkel is om die volumetriese waterinhoud van monsters by 5, 10 en 100 kPa en die grootte van die waterhouvermoë tussen 5 en 100 kPa en tussen 10 and 100 kPa te voorspel, kan baie waardevol wees. Tyd en geld kan potensieel bespaar word. Die modelle wat hoogs aanbevole is, is die modelle vir onversteurde monsters. Die modele is: By 5 kPa, VWlo = 0.47259 - 0.04712 rnedlumsand?" By 10 kPa, VWlo = 0.41292 - 0.04221 mediumsando.s By 100 kPa, VWlo = 0.48080 - 0.00254 fynsand - 0.0865 In mediumsand Tussen 5 en 100 kPa, WHVo = -29.523 + 3.394 fynsand Tussen 10 en 100 kPa, WHVo = -891.794 + 232.326 In fynsand + 38.006 In mediumsand

A field experiment was conducted to investigate the feasibility of making one measurement of soil moisture tension/content or crop water stress index (CWSI), and from it determine when and how much to irrigate. Cotton (Gossvpium hirsutum C.V. DPL-90) was planted in a sandy loam soil prepared with an underground drip irrigation system. Measurements of soil moisture tension, soil moisture content, plant temperature, and wet- and dry-bulb temperatures were made. The results indicate that soil moisture tension above 0.30 m and soil moisture content above 0.50 m in the root zone can be used as good indicators for CWSI or vice versa. Further work is required using different irrigation systems and different crops in larger land areas.

Master of Science
Department of Biological & Agricultural Engineering
Aleksey Y. Sheshukov
Soil erosion is one of the main reasons for agricultural land degradation in the world. Losses of land because of high soil erosion rates and rapidly expanding population result in significant reduction of cultivated land area per capita, and shortage of food on the global scale. Soil erosion can be a major source of sediment in the aquatic systems leading to reduction of organism population and poor water quality. Many factors affect soil erodibility, such as, soil properties, rainfall, topographic features, land use, and management practices, among others. The impacts of soil moisture content, however, are not well understood and. therefore, the primary goal of this study was to quantify two soil erodibility parameters, the erodibility coefficient and critical shear stress, under different soil moisture conditions using the jet erosion test (JET). The JET test uses the apparatus (called mini-JET) that creates an impinging jet of water into the soil and records the resulting scour depth over time. The scour depth time series are then fitted into a non-linear soil erosion equation, yielding the sought values of erodibility parameters. For this study, more than 40 soil samples were collected from several sites in Kansas, processed, and prepared to conduct JET tests in the lab setting. The effects of tillage and soil moisture content were of interest to this study. The results showed varied effects of soil type and sample soil moisture condition on the scour depth development and parameters sensitivity. The critical shear stress decreased and the erodibility coefficient increased with the increase of initial moisture content for clay loam soil, while critical shear stress did not change for sandy loam soil. The study also revealed higher erosive properties of soil collected from the tilled field compared to the no-till field.

Nowadays technology plays an important role in different sectors of farm management, in particular technologies based on soil moisture assessment have proven to be efficient in helping farmers who need timely techniques to determine crop water requirements. The main weakness of subsurface moisture sensing techniques is usually related both to the lack of cost-effectiveness of measurements and to unsuitable support scales with respect to the extension of the surface to be investigated. In this regard, Wireless Underground Sensor Network are increasingly used non-destructive tool specifically suitable for characterization and measurement. Wireless communication within a dense substance such as soil is, however, significantly more challenging than through air. This factor, combined with the necessity to conserve energy due to the difficulty of unearthing and recharging WUSN devices, requires that communication protocols be redesigned to be as efficient as possible. This work focuses on problem of implementing a Wireless Underground Sensor Network using LoRa technology, one of the most prominent Low Power Wide Area Network technologies, for communication. A system for monitoring soil moisture is implemented using low cost solutions and results are presented.

Meteorological measurements in the Walnut Gulch catchment in Arizona were used to synthesize a distributed, hourly-average time series of data across a 26.9 by 12.5 km area with a grid resolution of 480 m for a continuous 18-month period which included two seasons of monsoonal rainfall. Coupled surface-atmosphere model runs established the acceptability (for modeling purposes) of assuming uniformity in all meteorological variables other than rainfall. Rainfall was interpolated onto the grid from an array of 82 recording rain gauges. These meteorological data were used as forcing variables for an equivalent array of stand-alone Biosphere-Atmosphere Transfer Scheme (BATS) models to describe the evolution of soil moisture and surface energy fluxes in response to the prevalent, heterogeneous pattern of convective precipitation. The calculated area-average behavior was compared with that given by a single aggregate BATS simulation forced with area-average meteorological data. Heterogeneous rainfall gives rise to significant but partly compensating differences in the transpiration and the intercepted rainfall components of total evaporation during rain storms. However, the calculated area-average surface energy fluxes given by the two simulations in rain-free conditions with strong heterogeneity in soil moisture were always close to identical, a result which is independent of whether default or site-specific vegetation and soil parameters are used. Because the spatial variability in soil moisture throughout the catchment has the same order of magnitude as the amount of rain falling in a typical convective storm (commonly 10% of the vegetation's root zone saturation), in a semi-arid environment, any non-linearity in the relationship between transpiration and the soil moisture available to the vegetation has limited influence on area-average surface fluxes.

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.

Sufficient knowledge of near-surface soil moisture and sea surface salinity is essential for environmental studies. In recent years, efforts worldwide have focused on the use of imaging sensors onboard remote sensing satellites to capture such information. This thesis investigates if low-frequency radio signals originating from ground-based stations are suitable for the remote sensing of soil moisture and sea surface salinity. It is found that the time delay variation of 100 kHz Loran-C pulses may be representative of electrical conductivity changes along the propagation path, which have a direct link with soil and sea surface properties. The validation studies using Loran-C signals recorded at Bath and Harwich clearly show that the proposed method is capable of producing an estimation of soil moisture and sea surface salinity. This will form the basis for further validation and studies.

Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Civil and Environmental Engineering, 1996.
Includes bibliographical references (leaves 144-146).
by Eddie James Humphries, Jr.
M.S.

Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Civil and Environmental Engineering, 2001.
Includes bibliographical references (leaves 163-168).
This thesis addresses the question of how the early morning atmospheric thermodynamic structure affects the interaction between the soil moisture state and the growth and development of the boundary layer (BL), leading to the triggering of convection. It is concluded that in mid-latitudes, for matters of convective triggering and response to land surface conditions, the critical portion of the atmospher~approximately1 to 3 km above the ground surface is independent of geographic location and local synoptic setting. As long as the low levels of the troposphere are relatively humid but not extremely close to saturation, a negative feedback between soil moisture and rainfall is likely when the early morning temperature lapse rate in this region is dry adiabatic; a positive feedback is likely when it is moist adiabatic; and when there is a temperature inversion in this region, deep convection cannot occur, independent of the soil moisture. Additionally, when the low levels of the troposphere are extremely dry or very close to saturation, the occurrence of convection is determined solely by the atmospheric conditions. Essential characteristics of the temperature structure of the early-morning atmosphere are captured by a new thermodynamic measure, the Convective Triggering Potential (CTP), developed to distinguish between soundings favoring rainfall over dry soils from those favoring rainfall over wet soils. Many measures of atmospheric humidity are effective at separating atmospherically-controlled cases from cases where the land surface conditions can influence the likelihood for convection, but Hi low, a variation of a humidity index, proved most effective. A one-dimensional model of the planetary boundary layer (BL) and surface energy budget has been modified to allow the growing BL to entrain air from an observed atmospheric sounding. The model is used to analyze the impact of soil saturation on BL development and the triggering of convection in different atmospheric settings. Results from this 1D model and from the three-dimensional Fifth-Generation Penn State/NCAR Mesoscale Model (MM5) show a small but significant positive soil moisture-rainfall feedback in Illinois. This is consistent with an analysis of the distribution of early morning sounding values of CTP and Hi low from Illinois, though wind effects important in the MM5 simulations are not captured by the CTP-HIhow framework. From the MM5 simulations, it is concluded that the land surface condition can impact the potential for convection only when the atmosphere is not already predisposed to convect or not to convect. This atmospheric predisposition can be determined by analyzing the CTP, the Hi low, and the vertical profile of the winds. Analyses of Hi low scatter plots from radiosonde stations across the contiguous 48 United States reveal that positive feedbacks are likely in much of the eastern half of the country. The only area showing a potential negative feedback is in the Dryline and Monsoon Region of the arid southwest. Land surface conditions are unlikely to impact convective triggering in the rest of the western half of the country. Use of the lD BL model at four additional stations confirms that HilowTP-Hi low framework used in this nationwide analysis is valid for regions far removed from Illinois, where it was originally developed.
by Kirsten Lynn Findell.
Ph.D.

The thesis numerically models and investigates the effect of a variable water table on the soil moisture content. The modelling is done using COMSOL and Richards' equation. The temporal variation plots can be used to find the capillarity of the soil and its impact on other phenomenon such as vapor intrusion and infiltration.

Dans cette thèse, un algorithme de restitution à base de réseaux de neurones a été développé afin d’estimer l’humidité du sol à partir d’une combinaison d’observations satellitaires en micro-ondes, infrarouge et visible. Une estimation globale des valeurs mensuelles d’humidité du sol a été obtenue pour la période 1993-2000 et est fournie sur une grille à pixel de surface constante avec une résolution équatoriale de 0,25 ◦. Cette estimation de l’humidité du sol a été évaluée avec des données modélisées, des données de télédétec- tion et des observations in situ et a montré une bonne performance à différentes échelles spatiales et temporelles. Une analyse de contenu en information a montré que chacune des différentes observations satellites contribue à une information différente sur l’humidité du sol, avec les données micro-ondes actives plus sensibles à l’évolution temporelle et les données infrarouges thermiques reproduisant mieux les structures spatiales. En outre, une analyse de synergie a révélé que la combinaison de toutes les observations permet une réduction de l’incertitude de restitution de plus de 18 % et que la méthode des réseaux de neurones exploite de manière optimale la synergie des observations par comparaison avec autres approches. Une analyse a démontré la cohérence de l’humidité du sol resti- tuée avec d’autres produits satellitaires réprésentatifs d’autres paramètres hydrologiques (inondations, précipitations) à l’échelle du globe. Cela souligne le potentiel de nôtre jeu de données d’humidité du sol pour les études du cycle de l’eau terrestre. Enfin, il a été démon- tré que la méthode de réseaux de neurones proposée, constitue également un outil efficace pour évaluer les modèles de surface continentale ainsi que la modélisation des processus
In this thesis, a neural network based retrieval algorithm has been developed to compute surface soil moisture from a combination of microwave, infrared and visible satellite obser- vations. A global estimate of monthly mean soil moisture values has been computed for the period 1993-2000 and is provided on an equal-area grid with an equatorial resolution of 0. 25◦. This soil moisture estimate has been evaluated against modelled, remotely sensed and in situ observations and was found to perform well on different spatial and temporal scales. An information content showed that each of the various satellite observations con- tributes information about a different soil moisture variation, with the active microwave data being more sensitive to the temporal evolution and the thermal infrared data better capturing the spatial patterns. Furthermore, a synergy analysis revealed that the combina- tion of all observations permits a reduction of the retrieval uncertainty by more than 18% and that the neural network methodology optimally exploits the synergy of observations compared to other approaches. A joint analysis of various remotely sensed datasets of ter- restrial water cycle components demonstrated the coherence of the retrieved soil moisture with other retrieval products and with global hydrological processes. This underlined its potential to be used for observation-based studies of the terrestrial water cycle. Finally, it has been shown that the proposed neural network methodology also provides an effective tool to evaluate Earth System Models on both a variable and a process basis

Includes bibliographical references (leaves 106-118).
The research focuses on an objective analysis of austral summer rainfall variability over Zimbabwe as well as characterization of rainfall patterns and frequency analysis over southern Africa region. A statistical analysis of historical trends in climate extreme events is used and lays a foundation of projecting into future climates. A trend analysis done on rainfall patterns attained from SOMs approach compliments the RClimdex statistical approach and strengthens some of the historical trends findings on climate extremes. Thereafter, some exploratory research seeks to explain the trends observed using the land-atmosphere interactions and shows the response of rainfall to anomalous soil moisture conditions during an extreme wet and dry seasons using RegCM3. Finally, some radiation effects results are presented from these soil moisture perturbations experiments. Results show drying out patterns over the region from the historical records analysed. The trend analysis done with SOM arrays revealed a positive trend towards drier conditions and a negative trend for wet conditions. The climate extremes indices analysis complimented these findings as shown in the decrease in total precipitation and an increase in the number of dry spells. This is supported by the circulation patterns showing an increase in frequency of the 500hPa anticyclones and a decrease of low pressures. However, some high altitude stations showed an intensification of precipitation events. This would exacerbate need for proper planning of future water resource management and farming strategies. Soil moisture rainfall feedback mechanisms were not fully explored. However drier conditions experiments showed a stronger response to soil moisture perturbations than in wetter conditions experiments. No consistent response to soil moisture initialisation over southern Africa was found. The altitude does modulate these feedback mechanisms with low-lying areas depicting a stronger response. A better understanding of the observed rainfall patterns, historical climate trends and soil moisture-rainfall feedback mechanisms are essential for improved short-term and seasonal forecasting and will aid the generation of plausible climate change impact predictions.

The purpose of this work was to study and evaluate the effect of various mulches on the soil moisture content for sandy loam soil under high temperatures. The straw, wood shavings, and rock fragments were used with different rates of application for covering soil. Treatments were compared to bare soil. The moisture content was significantly higher for mulched treatments. However, the effectiveness of wood shavings and rock fragments mulches did not seem different in soil moisture content, while straw mulch did. The soil moisture content for the straw had the highest value at the end of the experiment. The straw mulch apparently conserved soil moisture over the other treatments. The conservation of soil moisture under 10, 15, 20 gm of straw mulch was occasioned by less direct contact soil surface with air surface and less possibility of heat transfer. The greatest savings in water was under 2.7 tons/ha of straw mulch.

This thesis introduces the implementation of different supervised learning techniques for producing accurate estimates of soil moisture content using empirical information, including meteorological and remotely sensed data. The models thus developed can be extended to be used by the personal remote sensing systems developed in the Center for Self-Organizing Intelligent Systems (CSOIS). The dfferent models employed extend over a wide range of machine-learning techniques starting from basic linear regression models through models based on Bayesian framework. Also, ensembling methods such as bagging and boosting are implemented on all models for considerable improvements in accuracy. The main research objective is to understand, compare, and analyze the mathematical backgrounds underlying and results obtained from dfferent models and the respective improvisation techniques employed.

Thesis (Ph.D.)--George Mason University, 2008.
Vita: p. 228. Thesis director: Menas Kafatos. Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Computational Sciences and Informatics. Title from PDF t.p. (viewed July 7, 2008). Includes bibliographical references (p. 210-227). Also issued in print.