RESUMEN
Built environment stocks have attracted much attention in recent decades because of their role in material and energy flows and environmental impacts. Spatially refined estimation of built environment stocks benefits city management, for example, in urban mining and resource circularity strategy making. Nighttime light (NTL) data sets are widely used and are regarded as high-resolution products in large-scale building stock research. However, some of their limitations, especially blooming/saturation effects, have hampered performance in estimating building stocks. In this study, we experimentally proposed and trained a convolution neural network (CNN)-based building stock estimation (CBuiSE) model and applied it to major Japanese metropolitan areas to estimate building stocks using NTL data. The results show that the CBuiSE model is capable of estimating building stocks at a relatively high resolution (approximately 830 m) and reflecting spatial distribution patterns, although the accuracy needs to be further improved to enhance the model performance. In addition, the CBuiSE model can effectively mitigate the overestimation of building stocks arising from the blooming effect of NTL. This study highlights the potential of NTL to provide a new research direction and serve as a cornerstone for future anthropogenic stock studies in the fields of sustainability and industrial ecology.
Asunto(s)
Entorno Construido , Aprendizaje Profundo , Ciudades , Industrias , JapónRESUMEN
The food-water-land-ecosystem (FWLE) nexus is fundamental for achieving sustainable development. This study examines the influence of urbanization on the FWLE nexus. Toward this end, land was deemed as an entry point. Therefore, the impact of urbanization on the nexus was explored based on changes in land use. We selected Shenzhen, a city in China, as the study area. First, a land change modeler was employed to analyze historical land-use changes from 2000 to 2010, to build transition potential submodels, and to project future land-use patterns for 2030 under a business-as-usual scenario. Second, based on land-use maps, we assessed habitat quality, water yield, and water supply from 2000 to 2030 using Integrated Valuation of Ecosystem Services and Tradeoffs. Moreover, crop production was estimated according to statistical materials. Finally, the study presents the analyses and discussion of the impacts of urbanization on ecosystem services related to the FWLE nexus. The results of land-use changes indicated that a significant expansion of artificial surfaces occurred in Shenzhen with varying degrees of decrease in cultivated land, forest, and grassland. Furthermore, habitat quality, water supply, and crop production decreased evidently due to rapid urbanization. In contrast, the total water yield indicated an upward trend owing to the increased water yield from increasing artificial surfaces, whereas water yield from other land-use areas declined, such as the forest and grassland. The results demonstrated a significant positive correlation between artificial surfaces and total water yield. However, negative correlations were observed in the interaction among habitat quality, water supply, and crop production. The study presented temporal and spatial assessments to provide an effective and convenient means of exploring the interactions and tradeoffs within the FWLE nexus, which, thus, contributed to the sustainable transformation of urbanization.
Asunto(s)
Ecosistema , Urbanización , China , Conservación de los Recursos Naturales , AguaRESUMEN
Sharing successful practices with other stakeholders is important for achieving SDGs. In this study, with a deep-learning natural language processing model, bidirectional encoder representations from transformers (BERT), the authors aimed to build (1) a classifier that enables semantic mapping of practices and issues in the SDGs context, (2) a visualizing method of SDGs nexus based on co-occurrence of goals (3) a matchmaking process between local issues and initiatives that may embody solutions. A data frame was built using documents published by official organizations and multi-labels corresponding to SDGs. A pretrained Japanese BERT model was fine-tuned on a multi-label text classification task, while nested cross-validation was conducted to optimize the hyperparameters and estimate cross-validation accuracy. A system was then developed to visualize the co-occurrence of SDGs and to couple the stakeholders by evaluating embedded vectors of local challenges and solutions. The paper concludes with a discussion of four future perspectives to improve the natural language processing system. This intelligent information system is expected to help stakeholders take action to achieve the sustainable development goals. Supplementary Information: The online version contains supplementary material available at 10.1007/s11625-022-01093-3.