Population spatialization with pixel-level attribute grading by considering scale mismatch issue in regression modeling

Mei, Yuao; Gui, Zhipeng; Wu, Jindi; Peng, Dehua; Li, Rui; Wu, Huayi; Wei, Zhengyang

doi:10.1080/10095020.2021.2021785

Cited by 12 publications

(10 citation statements)

References 59 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The differences among counties will not be easily averaged out. Moreover, our method is less affected by scale mismatch and can be transferred to cross-scale modeling 26 .…”

Section: Methodological Frameworkmentioning

confidence: 99%

“…(iii) Validating the accuracy of the methods. The performance of the grazing spatialization model was evaluated through a comparison of the predicted value with census value 26 . Accuracy validation indexes, including coefficients of determination (R 2 ), root mean square error (RMSE), and mean absolute error (MAE), were used to evaluate the performances of the proposed RF-based models (Table 2), as presented in Eq.…”

Section: Methodological Frameworkmentioning

confidence: 99%

“…The combined approach can integrate the advantages of the two data production methods. The CSFs extraction method can narrow the scale differences in feature representation between administrative-level and pixel-level by replacing the average administrative-level value with a finer-level value, thereby increasing the training sample size and range and ameliorating the scale mismatch problem in the relationship transmission process 26 . Additionally, the random forest (RF) model, a widely used and well-understood non-parametric machine learning algorithm for regression analysis, is characterized by high computational efficiency and estimation accuracy, and has been used in the spatialization of various census data owing to its simplicity 25,27,28 .…”

Section: Background and Summarymentioning

confidence: 99%

See 2 more Smart Citations

A high-resolution gridded grazing dataset of grassland ecosystem on the Qinghai–Tibet Plateau in 1982–2015

Nan

Wang

et al. 2023

Sci Data

View full text Add to dashboard Cite

Grazing intensity, characterized by high spatial heterogeneity, is a vital parameter to accurately depict human disturbance and its effects on grassland ecosystems. Grazing census data provide useful county-scale information; however, they do not accurately delineate spatial heterogeneity within counties, and a high-resolution dataset is urgently needed. Therefore, we built a methodological framework combining the cross-scale feature extraction method and a random forest model to spatialize census data after fully considering four features affecting grazing, and produced a high-resolution gridded grazing dataset on the Qinghai–Tibet Plateau in 1982–2015. The proposed method (R2 = 0.80) exhibited 35.59% higher accuracy than the traditional method. Our dataset were highly consistent with census data (R2 of spatial accuracy = 0.96, NSE of temporal accuracy = 0.96) and field data (R2 of spatial accuracy = 0.77). Compared with public datasets, our dataset featured a higher temporal resolution (1982–2015) and spatial resolution (over two times higher). Thus, it has the potential to elucidate the spatiotemporal variation in human activities and guide the sustainable management of grassland ecosystem.

show abstract

“…The differences among counties will not be easily averaged out. Moreover, our method is less affected by scale mismatch and can be transferred to cross-scale modeling 26 .…”

Section: Methodological Frameworkmentioning

confidence: 99%

Section: Methodological Frameworkmentioning

confidence: 99%

Section: Background and Summarymentioning

confidence: 99%

See 1 more Smart Citation

A high-resolution gridded grazing dataset of grassland ecosystem on the Qinghai–Tibet Plateau in 1982–2015

Nan

Wang

et al. 2023

Sci Data

View full text Add to dashboard Cite

show abstract

“…The Random Forest model has the advantages of avoiding overfitting, having a high tolerance for outliers and noise, and measuring the importance of variables, and is widely used for population spatialization (Mei et al, 2022; Wang, Wang, et al, 2022). Since the spatiotemporal resolution of existing commonly used datasets, such as WorldPop at 100 m and 1 km from 2000 to 2020, LandScan Global Population Database at 1 km from 2000 to 2020, and Chinese population spatial distribution km grid dataset at 1 km from 1995 to 2019, cannot meet the requirements of this study.…”

Section: Methodsmentioning

confidence: 99%

Guiding and constraining reclamation for coastal zone through identification of response thresholds for ecosystem services supply–demand relationships

Liu,

Yan,

Han

et al. 2024

Land Degrad Dev

View full text Add to dashboard Cite

Enhancing the understanding and management of ecosystem services (ESs) supply and demand under the influence of global change is crucial for sustaining human well‐being and promoting sustainable development. However, the comprehension of changes in ecosystem service dynamics in the face of complex human disturbances remains limited. This study focuses on four representative ESs in the northern Liaodong Bay, a typical global estuary wetland, to elucidate alterations in the supply and demand of these ESs and comprehend the impact mechanisms arising from reclamation activities. The findings reveal the following key insights: between 1993 and 2020, the supply and demand for water yield disparity became increasingly pronounced, while the equilibrium between supply and demand for carbon storage shifted from surplus to balance. In contrast, the excess supply–demand of food supply and landscape esthetics remained relatively consistent. This study delves deeper into the non‐linear relationship between reclamation activities and ESs supply–demand relationships, with each ecosystem service exhibiting distinct patterns. As reclamation intensity rises, the supply–demand ratio of ESs exhibits varying trends: food supply and carbon storage demonstrate an inverted U‐shaped relationship, water yield follows a U‐shaped pattern, and landscape esthetics progressively decline. By identifying reclamation thresholds, different management strategies should be applied to areas exceeding and falling below these thresholds, with constraints and guidance being necessary for the former and the latter, respectively. These findings advance our understanding of ESs dynamics in the context of human‐driven changes and provide practical implications for sustainable ecosystem management in coastal areas.

show abstract

“…Numerous studies have already explored techniques for disaggregating population estimates to fine-grained grid cells (Langford, 2006;Mei et al, 2022;Mennis, 2009;Silva et al, 2013) as a way of addressing this problem.…”

Section: Introductionmentioning

confidence: 99%

Modeling population distribution: A visual and quantitative analysis of gradient boosting and deep learning models for multi‐output spatial disaggregation

Georgati,

Monteiro,

Martins

et al. 2024

Transactions in GIS

View full text Add to dashboard Cite

Spatially aggregated data on socio‐demographic groups often fail to capture the population's spatial heterogeneity in cities. This poses challenges for urban planning, particularly when addressing the needs of groups such as migrants or families with children. Moreover, the commonly provided aggregated units, such as census tracts, vary in size and across data sources. Existing literature on disaggregation typically handles individual subgroups separately, ignoring their interrelations in the downscaling process. This article explores the potentials of multi‐output regression models for simultaneous spatial downscaling of multiple groups and conducts a detailed spatial error analysis using individualized neighborhoods. We experiment with self‐training gradient‐boosting trees and fully convolutional neural networks, assessing the quality of results against ground truth data at the target resolution. We show that the evaluation of the disaggregated results at this detailed resolution requires unconventional methods. The methodology proves convenient and achieves high‐accuracy results using input datasets of building features.

show abstract

Population spatialization with pixel-level attribute grading by considering scale mismatch issue in regression modeling

Cited by 12 publications

References 59 publications

A high-resolution gridded grazing dataset of grassland ecosystem on the Qinghai–Tibet Plateau in 1982–2015

A high-resolution gridded grazing dataset of grassland ecosystem on the Qinghai–Tibet Plateau in 1982–2015

Guiding and constraining reclamation for coastal zone through identification of response thresholds for ecosystem services supply–demand relationships

Modeling population distribution: A visual and quantitative analysis of gradient boosting and deep learning models for multi‐output spatial disaggregation

Contact Info

Product

Resources

About