京都大学 / 0048 / 新制・課程博士 / 博士(農学) / 甲第20816号 / 農博第2256号 / 新制||農||1055(附属図書館) / 学位論文||H30||N5098(農学部図書室) / 京都大学大学院農学研究科森林科学専攻 / (主査)教授 神﨑 護, 教授 北島 薫, 教授 德地 直子 / 学位規則第4条第1項該当 / Doctor of Agricultural Science / Kyoto University / DGAM

No description available.

Deforestation of tropical forests is one of the most pressing environmental problems of the twenty-first century, leading to the loss of environmental services such as climate regulation and biodiversity. The expansion of the agricultural frontier by small landholder farmers continues to be one of the major drivers of land use change in the Amazon region. Much of the recent research in the Brazilian Amazon has been focused on modeling their behavior in order to prescribe policies that can curb current deforestation rates and promote more sustainable land use practices. The availability of more sophisticated remote sensing and economic modeling tools has led to the proliferation of agricultural household level models that attempt to explain land use change processes at the farm level. This dissertation tests the household life cycle theory in one of the oldest colonization fronts in the Brazilian Amazon: Rondônia, now a post-frontier. The study examines household and farm level changes over time for specific aspects of the frontier process that can be tested using the household life cycle theory. This study introduces important additions to the life cycle theory in order to consider the more dynamic and complex set of factors that characterize modern frontier processes. Specifically the study examines: (1) property fragmentation and expansion processes, (2) property ownership, turnover and change, and (3) land use change processes at the property level. These are linked to changes in the social and economic features of the smallholder farmer as it moves along its life cycle. The central hypothesis is that these changes in property and land use dynamics can be explained by the corresponding changes in the life cycle of the household as the frontier evolves over time into a post-frontier. It was found that the household life cycle theory did not adequately explain land use change processes over time. As the frontier evolved into the modern post-frontier, the labor and drudgery constraints associated with the initial frontier processes, as exemplified in the household life cycle theory, became less relevant. The Sauerian concept of cultural successions and the concept of scale from hierarchical ecology are used in order to explain the apparent inconsistencies found between the household life cycle theory and land use change processes over time and at different scales of analysis. The household life cycle theory is a useful theoretical framework from which to examine the effects of household level factors on land use; however, this must be embedded within concepts of time and scale that determine their differentiated impact and behavior. Existing plans to expand road infrastructure into the Amazon region will open-up previously inaccessible rainforest regions to agricultural frontier expansion at a scale unprecedented since the mid-eighties. Findings from this study reveal that policies based on household life cycle postulates will have limited impacts in reducing deforestation rates and promoting sustainable land use practices. Appropriate accounting of the social and environmental costs of future infrastructure development projects should consider associated frontier agricultural expansion costs to discourage further deforestation. / Ph. D.

Characterizing streamflow trends is important for water resources management. Streamflow conditions, and trends thereof, are critical drivers of all aspects of stream geomorphology, sediment and nutrient transport, and ecological processes. Using the non-parametric modified Mann-Kendall test, we analyzed streamflow trends from 1996 to 2022 for the Southern Appalachian (SA) region of the U.S. The forested uplands of the SA receive high amounts of rain and act as a "water tower" for the surrounding lowland area, both of which have experienced higher than average population growth and urban development. For the total of 127 USGS gages with continuous streamflow measurements, we also evaluated precipitation and land change rates and patterns within the upstream contributing areas. Statistical methods (i.e., generalized linear models) were then used to assess any linkages between land cover change (LCC) and streamflow trends. Our results show that 42 drainage areas are experiencing increasing trends in their precipitation, and 1 is experiencing a negative trend. A total of 71 drainage areas are experiencing increasing trends in either their annual streamflow minimums, maximums, medians, or variability, with some experiencing changes in multiple. From our models, it is suggested that agricultural expansion is associated with increasing minimum streamflow trends, but increasing precipitation is also positively linked. With this information, water managers would be aware of which areas are experiencing changes in streamflow amounts from LCC or precipitation and could then apply this in planning and predictions. / Master of Science / Water availability is important for resources management. Streamflow is a measure of available surface water and is an important component in the hydrological cycle. Using the non-parametric modified Mann-Kendall test, we analyzed streamflow trends from 1996 to 2022 for the Southern Appalachian (SA) region of the U.S. The forested uplands of the SA receive high amounts of rain and act as a "water tower" for the surrounding lowland area, both of which have experienced higher than average population growth and city expansion. For the total of 127 USGS gages with continuous streamflow measurements, we also evaluated precipitation and land cover change rates within the area upstream of the gage (or drainage/contributing area). Statistical methods (i.e., generalized linear models) were then used to assess any linkages between land cover change (LCC) and streamflow trends. Our results show that 42 drainage areas are experiencing increasing trends in their precipitation, and 1 is experiencing a negative trend. A total of 71 drainage areas are experiencing increasing trends in either their annual streamflow minimums, maximums, medians, or variability, with some experiencing changes in multiple. From our models, it is suggested that agricultural expansion is associated with increasing minimum streamflow trends, but increasing precipitation is also positively linked. With this information, water managers would be aware of which areas are experiencing changes in streamflow amounts from LCC or precipitation and could then apply this in planning and predictions.

This study compares suspended sediment and land use/land cover in the watershed of Weeks Bay, Alabama. Using Landsat thematic mapper imagery, potential high and low erosion sites were determined based on the increase in urban development form 2002 to 2005. In situ sediment sampling was used to test the hypothesis that the high erosion potential sites have larger amounts of suspended sediments. Additionally, sampling was performed along the Fish and Magnolia rivers to establish a background total suspended sediment level. The background study established an average total suspended sediment concentration of 18.71 mg/L for the Fish River and 17.47 mg/L for the Magnolia River, which are higher than previous studies. The results of the comparison between suspended sediments and land use/land cover proved to be more complex than expected due to variation in precipitation, to the 30 m satellite resolution, and to the criteria for classifying urban land use.

No description available.

No description available.

No description available.

In recent decades land use and land cover change (LULCC) has occurred throughout the Intermountain West. The Crown of the Continent Ecosystem (CCE) extends along the Rocky Mountains adjacent to the Canada-U.S. International border. In the U.S. portion of the CCE, located in northwestern Montana, development has increased since the 1990s, largely because of urban to rural migration. The CCE has become an amenity-based destination, which in turn is likely to threaten its terrestrial and aquatic ecological diversity (Quinn and Broberg 2007). Specifically, development pressures on private lands surrounding federally protected lands, are intensifying and thus threatening core habitat of native species and connectivity of forested areas. By characterizing the spatial and temporal patterns of LULCC, we can better understand landscape-scale changes influenced by human-environment interactions. Using National Land Cover Database (NLCD) products, I identified areas that have experienced land cover change for three time periods: 1992-2001, 2001-2006, and 2006-2011. Additionally, I used case studies to further investigate LULCC in the study area. The findings suggest that the highest rates of development in proximity to Glacier National Park were dependent on existing urban land cover, meaning existing roadway infrastructure and established urban areas saw the greatest urban development. Additionally, communities adjacent to Glacier National Park were hotspots for urban development. Based on the results, areas in proximity to federally protected lands are likely to experience continued urban intensification over the next few decades. / Master of Science

Invasive plant species threaten native plant communities and inhibit efforts to restore disturbed landscapes. Surface coal mines in the Appalachian Mountains are some of the most disturbed landscapes in North America. Moreover, there is not a comprehensive understanding of the land cover characteristics of post- mined lands in Appalachia. Better information on mined lands' vegetative cover and ecosystem recovery status is necessary for implementation of effective environmental management practices. The invasive autumn olive (Elaeagnus umbellata) is abundant on former coal surface mines, often outcompeting native trees due to its faster growth rate. The frequent revisit time and spatial and spectral resolution of Landsat satellites make Landsat imagery well suited for mapping and characterizing land cover and forest recovery on former coal surface mines. I performed a multitemporal classification using a random forest analysis to map autumn olive on former and current surface coal mines in southwest Virginia. Imagery from the Operational Land Imager on Landsat 8 were used as input data for the study. Calibration and validation data for use in model development were obtained using high-resolution aerial imagery. Results indicate that autumn olive cover is sufficiently dense to enable detection using Landsat imagery on approximately 12.6% of the current and former surface coal mines located in the study area that have been mined since the early 1980s. The classified map produced here had a user's and producer's accuracy of 85.3% and 78.6% respectively for the autumn olive coverage class. Overall accuracy in reference to an independent validation dataset was 96.8%. These results indicate that autumn olive growing on reclaimed coal mines in Virginia and elsewhere in the Appalachian coalfields can be mapped using Landsat imagery. Additionally, autumn olive occurrence is a significant landscape feature on former surface coal mines in the Virginia coalfields. / Master of Science