The United States Army land managers are facing a difficult task of balancing environmental quality and military land carrying capacity when planning missions. The increase in soil erosion and landscape fragmentation caused by intensive military training degrades environmental quality and restricts military missions simultaneously. So far, no effective tools can be applied to quantitatively assess the environmental quality of military training facilities. This study aims at overcoming the existing gaps in land management of the U.S. Army installations. In this study, spatial metrics were selected and used to quantify landscape quality and further their correlations with landscape aesthetics indicators were investigated to seek surrogates of the immeasurable indicators. The spatial metrics were then combined with other environmental variables including soil erosion, water quality, and noise to create an integrated indicator that comprehensively measures environmental quality for the U.S. Army installations using spatial multi-criteria decision analysis. The methodology proposed in this study was tested at Fort Riley Installation, Kansas. The obtained important results included i) Landsat Thematic Mapper TM imagery was better at identifying land cover categories than India Remote Sensing Imagery and their Brovey transformations and Principal Component Analysis (PCA); ii) Too fine of a spatial resolution of imagery led to a great number of small patches and degraded the accuracy of landscape segmentation; iii) both landscape shape index (LSI) and Aggregation Index (AI) had statistically significant correlation with military training intensity and quantified the landscape fragmentation well along with both LSI and AI had a significant negative correlation; iv) there were moderate correlations of LSI and AI with landscape complexity and Interspersion and Juxtaposition index (IJI) with disturbance; v) the landscape level environmental quality indicator obtained comprehensively and well quantified the overall environmental health and its dynamics, while the patch level indicator detailed the local environmental quality. The significant contributions made in this study included i) exploring the relationships of landscape aesthetic evaluations with spatial metrics variables and further incorporating the spatial metrics as surrogates of the landscape aesthetic evaluations into derivation of comprehensive environmental quality indicator; and ii) developing a practical method to integrate the individual factors into a comprehensive environmental quality indicator at both landscape and patch levels based on sustainable environmental health and military land carrying capacity. Without doubt, this study can provide effective tools for the Army land managers to accurately assess environmental quality and effectively plan military training on the installations. It is also expected this methodology can be applied to management of other lands such as agricultural, forested, and industrial lands, etc.

Master of Science / Department of Statistics / Perla E. Reyes Cuellar / Data with ordinal responses may be encountered in many research fields, such as social, medical, agriculture or financial sciences. In this paper, we present a case study on cumulative logit models with low frequency and mixed effects and discuss some strengths and limitations of the current methodology. Two plant pathologists requested our statistical advice to fit a cumulative logit mixed model seeking for the effect of six commercial products on the control of a seed and seedling disease in soybeans in vitro. In their attempt to estimate the model parameters using a generalized linear mixed model approach with PROC GLIMMIX, the model failed to converge. Three alternative approaches to solve the problem were examined: 1) stratifying the data searching for the random effect; 2) assuming the random effect would be small and reducing the model to a fixed model; and 3) combining the original categories of the response variable to a lower number of categories. In addition, we conducted a power analysis to evaluate the required sample size to detect treatment differences. The results of all the proposed solutions were similar. Collapsing categories for a cumulative/proportional odds model has little effect on estimation. The sample size used in the case study is enough to detect a large shift of frequencies between categories, but not for moderated changes. Moreover, we do not have enough information to estimate a random effect. Even when it is present, the results regarding the fixed factors: pathogen, evaluation day, and treatment effects are the same as the obtained by the fixed model alternatives. All six products had a significant effect in slowing the effect of the pathogen, but the effects vary between pathogen species and assessment timing or date.

The cumulative effects of watershed development and large water withdrawals are placing the sustainability of freshwater resources at risk due to alteration of watershed hydrology, stream geomorphology, groundwater recharge, and adverse effects to the aquatic ecology of water resources. The consideration of cumulative environmental effects in development decisions under current project-specific assessment does not fully encompass the interacting effects of multiple stressors over space and time. As a result, the cumulative effects of land uses and development on watershed processes are not properly assessed and managed. There is a recognized need to shift from local, project-scale cumulative effects assessments to broader, landscape, or regional scale assessments to accurately assess cumulative effects to watershed processes and river system condition. The problem is that there is little understanding of the current capacity to do so. This research: i) developed a set of indicators for evaluation of regional capacity to support watershed cumulative effects assessment and management (CEAM) requisites, ii) applied those indicators to the South Saskatchewan Watershed (SSW), iii) identified capacity needs and constraints to watershed CEAM in SSW, and iv) identified lessons learned and opportunities for capacity building to support watershed CEAM principles and practice. Capacity indicator questions were developed for a set of eight institutional requirements for watershed CEAM, identified from a previous study of watershed CEAM in the SSW. Research methods included a web-based survey of academics, regulators, industry and environmental organizations, which consisted of both closed ended and open-ended questions based on the capacity indicators. Survey results were analyzed using the Statistical Package for the Social Sciences and qualitative methods. Results indicate that the primary threats to water quality and quantity in the SSW, as identified by study participants, are broad-scale stressors that are not subject to project-specific environmental assessment regulations. To address these broad-scale stresses, cumulative effects assessment at the regional level needs to be done; however, it was identified that there is currently a lack of mechanisms to support watershed CEAM. The need for a lead agency, multi-stakeholder collaboration, and financial and human resources were identified as the most important requisites from the research results for implementing and sustaining watershed CEAM programs. Research results revealed that watershed CEAM cannot be driven solely ‘bottom-up’ and government must lead watershed CEAM activities. Participants noted that there is collaboration ongoing in the SSW to meet CEAM objectives, but it is limited. There is a lack of clarity around common goals for watershed and sub-watershed management, and a lack of transparency in sharing data. Many participants commented that expertise is available for watershed CEAM, but there is a lack of organizational and financial resources to develop successful plans and actions.

This study investigates the degree to which the cumulative risk index predicted school readiness in a Head Start population. In general, the reviewed studies indicated the cumulative risk model was efficacious in predicting adverse developmental outcomes. This study built on this literature by investigating how child, parent, and family risk factors predicted school readiness in Head Start children using two statistical models. Specific aims of this study included identifying 1) to what degree multiple predictors contributed to school readiness and 2) to what degree the cumulative risk index contributed to school readiness. Participants included 176 Head Start children ages 3 to 5 years. Data were analyzed using multivariate regression to determine if the cumulative risk model was a stronger predictor of school readiness than any risk factor in isolation. Hierarchical regression was also utilized to determine if individual risk factors contributed anything above and beyond the sum, the cumulative risk index. Multiple regression analysis revealed that older age and previous enrollment in Head Start predicted higher scores, while low income predicted lower scores, as did taking the test in Spanish. Analysis also revealed that higher scores on the cumulative risk index predicted lower test scores. The analysis revealed that the individual risk factors did not contribute to the model above and beyond the cumulative risk index. Adding the individual risk factors did not account for more variance than using gender, age, and the cumulative risk index as the only predictors. Similarly, the cumulative risk index did not account for more variance than using age and gender as the only predictors. The current study adds empirical support to the continued use of the cumulative risk model in predicting adverse developmental outcomes.

none

The low back loading to which an individual is exposed has been linked to injury and the reporting of low back pain. Despite extensive research on the spine and workplace loading exposures, statistics indicate that efforts to date have not led to large reductions in the reporting of these injuries. One possible cause for the apparent ineffectiveness of interventions may be a poorly defined understanding of the mechanical exposures of the spine during work related activities. There are sophisticated models that can predict spine loads and are responsive to how an individual moves and uses their muscles, however the models are complex and require extensive data collection to be implemented. This fact has prevented these models from being employed in industrial settings and the simplified surrogate methods that are being employed may not be predicting load exposures well. Therefore, this work focused on examining surrogate methods that can produce estimates of spine loading equal to our most complex laboratory based models. In addition, our understanding of spine tolerance to combined motion and load has been based upon in-vitro work that has not accurately represented coupled physiologic compression and flexion or has not investigated potentially beneficial loading scenarios. The result has been a lack of clear data indicating when motion should be treated as the primary influence in injury development or when load is the likely injury causing exposure. As a result, research was conducted to determine the interplay between load and motion in cumulative injury development, as well as investigating the potential of static rest periods in mitigating the effects of cumulative compression. Study one examined the potential utility of artificial neural networks as a data reduction approach in obtaining estimates of time-varying loads and moments equal in magnitude to those of EMG-assisted and rigid link models. It was found that the neural network approach under predicted peak force and moment exposures, but produced strong predictions of average and cumulative exposures. Therefore this method may be a viable approach to document cumulative loads in industrial settings. Study two compared the load and moment estimates from a currently employed, posture match based ergonomic assessment tool (3DMatch) to those obtained with an EMG-assisted model and those predicted with a rigid link modeling approach. The results indicated that 3DMatch over predicted peak moments and cumulative compression. However, simple correction approaches were developed which can adjust the predictions to obtain more physiologic estimates. Study three employed flexion/extension motion with repetitive compression loading profiles in an in-vitro study, with both load and motion profiles being obtained from measures in study 1. It was found that at loads above 30% of a spine’s compressive tolerance, repetitive flexion/extension would not lead to intervertebral disc injury prior to an endplate or vertebral fracture occurring. However, as loads fall below 30% the likelihood of experiencing a herniation increases, while the overall likelihood of an injury occurring decreases. Comparison to relevant studies indicated that while repetitive flexion did not alter the site of injury it appeared to degrade the ability of the spine to tolerate compression. Finally, study four employed dynamic compression while the spine was maintained in a neutral posture to investigate the effects of ‘rest’, or periods of static low level loading, on altering the amount of load tolerated prior to injury. It was found that there was a non-linear relationship between load magnitude and compressive tolerance, with increasing load magnitude exposures leading to decreasing cumulative load tolerances. Periods of low level static loading did not alter the resistance of the spinal unit to cumulative compression or impact the number of cycles tolerated to failure. In summary, this work has examined methods that may allow for better predictions of spine loading in the workplace without the large data demands of sophisticated laboratory approaches. Where possible, suggestions for optimal implementation of these surrogates have been developed. Additionally, in-vitro work has indicated a load threshold of 30%, above which herniation is not likely to occur during dynamic repetitive loading. Furthermore, the insertion of static rest periods into dynamic loading scenarios did not improve the spine’s failure tolerance to loading, indicating that care should be exercised when determining optimal loading paradigms. In combination, the applied methods that have been developed and the information regarding injury development that has been obtained will help to refine our understanding of the exposures and tolerances that define mechanical injury in the spine.

The low back loading to which an individual is exposed has been linked to injury and the reporting of low back pain. Despite extensive research on the spine and workplace loading exposures, statistics indicate that efforts to date have not led to large reductions in the reporting of these injuries. One possible cause for the apparent ineffectiveness of interventions may be a poorly defined understanding of the mechanical exposures of the spine during work related activities. There are sophisticated models that can predict spine loads and are responsive to how an individual moves and uses their muscles, however the models are complex and require extensive data collection to be implemented. This fact has prevented these models from being employed in industrial settings and the simplified surrogate methods that are being employed may not be predicting load exposures well. Therefore, this work focused on examining surrogate methods that can produce estimates of spine loading equal to our most complex laboratory based models. In addition, our understanding of spine tolerance to combined motion and load has been based upon in-vitro work that has not accurately represented coupled physiologic compression and flexion or has not investigated potentially beneficial loading scenarios. The result has been a lack of clear data indicating when motion should be treated as the primary influence in injury development or when load is the likely injury causing exposure. As a result, research was conducted to determine the interplay between load and motion in cumulative injury development, as well as investigating the potential of static rest periods in mitigating the effects of cumulative compression. Study one examined the potential utility of artificial neural networks as a data reduction approach in obtaining estimates of time-varying loads and moments equal in magnitude to those of EMG-assisted and rigid link models. It was found that the neural network approach under predicted peak force and moment exposures, but produced strong predictions of average and cumulative exposures. Therefore this method may be a viable approach to document cumulative loads in industrial settings. Study two compared the load and moment estimates from a currently employed, posture match based ergonomic assessment tool (3DMatch) to those obtained with an EMG-assisted model and those predicted with a rigid link modeling approach. The results indicated that 3DMatch over predicted peak moments and cumulative compression. However, simple correction approaches were developed which can adjust the predictions to obtain more physiologic estimates. Study three employed flexion/extension motion with repetitive compression loading profiles in an in-vitro study, with both load and motion profiles being obtained from measures in study 1. It was found that at loads above 30% of a spine’s compressive tolerance, repetitive flexion/extension would not lead to intervertebral disc injury prior to an endplate or vertebral fracture occurring. However, as loads fall below 30% the likelihood of experiencing a herniation increases, while the overall likelihood of an injury occurring decreases. Comparison to relevant studies indicated that while repetitive flexion did not alter the site of injury it appeared to degrade the ability of the spine to tolerate compression. Finally, study four employed dynamic compression while the spine was maintained in a neutral posture to investigate the effects of ‘rest’, or periods of static low level loading, on altering the amount of load tolerated prior to injury. It was found that there was a non-linear relationship between load magnitude and compressive tolerance, with increasing load magnitude exposures leading to decreasing cumulative load tolerances. Periods of low level static loading did not alter the resistance of the spinal unit to cumulative compression or impact the number of cycles tolerated to failure. In summary, this work has examined methods that may allow for better predictions of spine loading in the workplace without the large data demands of sophisticated laboratory approaches. Where possible, suggestions for optimal implementation of these surrogates have been developed. Additionally, in-vitro work has indicated a load threshold of 30%, above which herniation is not likely to occur during dynamic repetitive loading. Furthermore, the insertion of static rest periods into dynamic loading scenarios did not improve the spine’s failure tolerance to loading, indicating that care should be exercised when determining optimal loading paradigms. In combination, the applied methods that have been developed and the information regarding injury development that has been obtained will help to refine our understanding of the exposures and tolerances that define mechanical injury in the spine.

This study investigates the degree to which the cumulative risk index predicted school readiness in a Head Start population. In general, the reviewed studies indicated the cumulative risk model was efficacious in predicting adverse developmental outcomes. This study built on this literature by investigating how child, parent, and family risk factors predicted school readiness in Head Start children using two statistical models. Specific aims of this study included identifying 1) to what degree multiple predictors contributed to school readiness and 2) to what degree the cumulative risk index contributed to school readiness. Participants included 176 Head Start children ages 3 to 5 years. Data were analyzed using multivariate regression to determine if the cumulative risk model was a stronger predictor of school readiness than any risk factor in isolation. Hierarchical regression was also utilized to determine if individual risk factors contributed anything above and beyond the sum, the cumulative risk index. Multiple regression analysis revealed that older age and previous enrollment in Head Start predicted higher scores, while low income predicted lower scores, as did taking the test in Spanish. Analysis also revealed that higher scores on the cumulative risk index predicted lower test scores. The analysis revealed that the individual risk factors did not contribute to the model above and beyond the cumulative risk index. Adding the individual risk factors did not account for more variance than using gender, age, and the cumulative risk index as the only predictors. Similarly, the cumulative risk index did not account for more variance than using age and gender as the only predictors. The current study adds empirical support to the continued use of the cumulative risk model in predicting adverse developmental outcomes.

Knowledge of the relationship between rate of nitrification and nitrous oxide (N2O) emission, and between cumulative nitrification and N2O emission is important for developing N2O emission mitigation strategies. Gross nitrification and N2O from nitrification were determined using 15N labelling of inorganic N. N-Serve was added to delay nitrification and results showed an increase in rate of N2O emission with that of apparent nitrification in absence of N-Serve, but there was no relation in its presence. Same amount of cumulative N2O was emitted for same amount of nitrogen (N) apparently nitrified, regardless of N-Serve addition. There was no relation between N2O emission attributed to nitrification and gross nitrification with and without N-Serve. Again, same amount of cumulative N2O was emitted for same amount of gross nitrified N, regardless of N-Serve addition. These results imply that the amount of N nitrified dictates eventual cumulative N2O emitted, regardless of rate of nitrification. / February 2017

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.