the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
HiTIC-Monthly: A High Spatial Resolution (1 km×1 km) Monthly Human Thermal Index Collection over China from 2003 to 2020
- 1School of Geography and Planning, Sun Yat-sen University, Guangzhou 510006, China
- 2Institute of Environment, Energy and Sustainability, The Chinese University of Hong Kong, Hong Kong SAR, China
- 3School of Geospatial Engineering and Science, Sun Yat-sen University, Zhuhai 519082, China
- 4School of Management, Guangdong University of Technology, Guangzhou 510520, China
- 5Dalla Lana School of Public Hea lth, University of Toronto, Toronto, Ontario M5T 3M7, Canada
- 6School of Atmospheric Physics, Nanjing University of Information Science & Technology, Nanjing 210044, China
- 7Tianjin Municipal Meteorological Observatory, Tianjin 300074, China
- 8State Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing 100081, China
- 9State Key Laboratory of Severe Weather (LASW) and Key Laboratory of Atmospheric Chemistry (LAC), Chinese Academy of Meteorological Sciences, Beijing 100081, China
- 10Collegee of Resources and Environment, Fujian Agriculture and Forest University, Fuzhou 35002, China
- 1School of Geography and Planning, Sun Yat-sen University, Guangzhou 510006, China
- 2Institute of Environment, Energy and Sustainability, The Chinese University of Hong Kong, Hong Kong SAR, China
- 3School of Geospatial Engineering and Science, Sun Yat-sen University, Zhuhai 519082, China
- 4School of Management, Guangdong University of Technology, Guangzhou 510520, China
- 5Dalla Lana School of Public Hea lth, University of Toronto, Toronto, Ontario M5T 3M7, Canada
- 6School of Atmospheric Physics, Nanjing University of Information Science & Technology, Nanjing 210044, China
- 7Tianjin Municipal Meteorological Observatory, Tianjin 300074, China
- 8State Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing 100081, China
- 9State Key Laboratory of Severe Weather (LASW) and Key Laboratory of Atmospheric Chemistry (LAC), Chinese Academy of Meteorological Sciences, Beijing 100081, China
- 10Collegee of Resources and Environment, Fujian Agriculture and Forest University, Fuzhou 35002, China
Abstract. Human thermal comfort measures the combined effects of temperature, humidity, and wind speed, etc., and can be aggravated under the influences of global warming and local human activities. With the most rapid urbanization and the largest population, China is being severely threatened by aggravating human thermal stress. However, the variations of thermal stress in China at a fine scale have not been fully understood. This gap is mainly due to the lack of a high-resolution gridded dataset of human thermal indices. Here, we generate the first high spatial resolution (1 km1 km) dataset of monthly human thermal index collection (HiTIC-Monthly) over China from 2003 to 2020. In this collection, 12 commonly used thermal indicators are generated by the LGBM machine learning algorithm from multi-source gridded data, including MODIS land surface temperature, topography, land cover and land use, population density, and impervious surface fraction. Their accuracies were comprehensively assessed based on observations at 2419 weather stations across the mainland of China. The results show that our dataset has desirable performance, with mean R2, root mean square error, mean absolute error, and bias of 0.996, 0.693 °C, 0.512 °C, and 0.003 °C, respectively, by averaging the 12 indicators. Moreover, the predictions exhibit high agreements with observations across spatial and temporal dimensions, demonstrating the broad applicability of our dataset. The comparison with two existing datasets also suggests that our high-resolution dataset can describe a more explicit spatial distribution of the thermal information, showing great potentials in fine-scale (e.g., intra-urban) study. Further investigation reveals that nearly all indicators exhibit increasing trends in most parts of China during the year 2003~2020. The increase is especially stronger in North China, Southwest China, the Tibetan Plateau, and parts of Northwest China, and in the spring and summer seasons. The HiTIC-Monthly dataset is publicly available via https://zenodo.org/record/6895533 (Zhang et al., 2022a).
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(3526 KB)
-
Supplement
(2228 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(3526 KB) -
Supplement
(2228 KB) - BibTeX
- EndNote
Journal article(s) based on this preprint
Hui Zhang et al.
Interactive discussion
Status: closed
-
RC1: 'Comment on essd-2022-257', Minyan Wang, 23 Aug 2022
Major comments.
The dataset is very import to study the climate change during this period over this region. The high spatial resolution and the monthly temporal resolution are basically reasonable, based on the previous datasets. The use of homogenous meteorological observation data is the right way to evaluate. This research is novel and belongs to the application of big data for large volume of Satellite data used, and machine learning algorithm. The authors are from 8 universities/colleges. The results supply important reference for other research fields like the field of geoscience, biology, sociology, and medicine and engineering.
- Line 415, harm, it is suggested to change the word. All 11 indices are calculated based on SAT, so the accuracy of this dataset is up to the accuracy of the dataset (Zhang 2022b) and the algorithm. SAT has important influence to the other 11 indices. How to understand these?
- (Zhang 2022b) SAT is LST, but not Tair (1.5 meters above the surface). Why use LST but not Tair in Table 1 to compute the other 11 indices?
- Line 157-158 “Section 6 compares our products with two existing datasets, and the main findings of this paper are summarized in Section 7”. Where is section 6 and section 7? They are data availability and conclusions.
- Why not keep Figures S1-S9, Tables S1-S5 as as formal ones? Please consider again which to keep in the manuscript.
- How to get monthly data from MODIS daily mid-daytime 13:30 and mid-nighttime 01:30 LST? The monthly mean is different considering of the diurnal variation and satellite observations from ascending orbits and descending orbits.
- Line 190-191, how to compute and use the covariates?
- HiTIC? What is i? Is it HTI (human thermal index)? No “I” after “human” and before ” thermal”.
Minor comments.
- Add abbr. of LGBM when it is first used. Light Gradient Boosting Machine. Check others, please.
- Author contribution. It seems only H.Z is responsible for analysis and data processing, others are all involved in writing. For computations, are there anyone else in the author list?
- Line 439. The unit of the dataset is degree or 0.01 degree? Is 0.01 the scale factor?
- Please revise the followings.
- Line 247-248, as the figure displays
- Line 384, Figures S8j&m. “Figure S8”, not figures, and add a blank before j.
- It is suggested to delete equations (1-2) and (3-6), add the reference.
- If necessary, please keep the same description of the time period for there are many kinds in the manuscript, like 2003~2020, from 2003 to 2020, during the year 2003~2020, (2003 to 2020), during 2003~2020, from January 2003 to December 2020, 2003-2020.
- Line 755. Delete the last sentence. Line 759, change : to . after 2003~2020.
- Add longitude and latitude (or the numerical scope signs) in Fig 1, 4-7, 9-12, 14.
- Figure 8. Prediction accuracies of 12 human thermal indices… Add 12. In individual years, change the description. Time series? Add the description of (a) (b) (c).
- Figure 10. Spatial distributions of 12 human thermal indices... Add 12.
- Line 780. Figure 11. National average, are you sure to write like this? Check it in other places in the manuscript. Just use average is fine. Delete “straight”. Line 789, Figure 13, national, delete this. Add 12, 12 human thermal indices.
- Line 785, Figure 12, add 12. “the trends of annual mean”, please consider how to express better. Inter-annual variation?
- Line 795, Figure 14. HITIC, is it “i”? in mainland China, is it over mainland of China? Please check the description of “mainland China” in the manuscript. In July 2018, move to the end of four major UAs. Delete i.e., change to :.
- AC1: 'Reply on RC1', Ming Luo, 01 Oct 2022
- AC4: 'Reply on RC1', Ming Luo, 17 Nov 2022
-
RC2: 'Comment on essd-2022-257', Anonymous Referee #2, 28 Aug 2022
The authors have produced a high-resolution (1 km×1 km) thermal index collection at a monthly scale (HiTIC-Monthly) in China during 2003 to 2020, with 12 widely used human thermal indices. The authors have created a high-resolution products for quantifying thermal index in China, which is valuable to the scientific community. I have some comments to be addressed by the authors.
1. The biggest concern is the temporal resolution. Why do the authors choose monthly resolution, rather than daily? Daily products would be extremely useful to characterize extreme events, which are of societal importance.
2. Lines 168-169: How about the impacts of precipitation on thermal indices? Have the authors considered precipitation as a covariate?
3. Figures 7 and 8: The results indicate spatial variability of bias in the thermal indices. What factors drive the spatial variability of the bias? Meanwhile, there is also temporal variability in the bias (Figure 8), and what is the drivers of this variability? Are the spatial and temporal variabilities of the bias related to background climates?
4. One way of evaluating the quality of these products is to evaluate the EOFs of these products. For example, what are the first three EOFs in each product? How do the temporal coefficients change over time across these products? Such spatial-temporal evaluation would be desirable.- AC2: 'Reply on RC2', Ming Luo, 01 Oct 2022
- AC5: 'Reply on RC2', Ming Luo, 17 Nov 2022
-
RC3: 'Comment on essd-2022-257', Minyan Wang, 26 Oct 2022
3) Line 157-158, suggest: Comparisons on our products with two existing datasets are in Section 5, data availability is provided in Section 6, ...
The 3rd response to the reviewer2 (spatial variability, temporal variability, background climates), Please consider how to answer the question directly.
-
AC3: 'Reply on RC3', Ming Luo, 07 Nov 2022
1.Line 157-158, suggest: Comparisons on our products with two existing datasets are in Section 5, data availability is provided in Section 6, ...
Response: Revised per your suggestion.
2.The 3rd response to the reviewer2 (spatial variability, temporal variability, background climates), Please consider how to answer the question directly.
Response: Thank you very much for your comment. Below please see our updated and direct responses:
As shown in Figure 7, the biases exhibit zonal variations across the space, i.e., positive bias values tend to distribute in northern China and negative values are mainly located in the south. This spatial variability is likely caused by the generally low and high temperatures in the north and south, respectively. The extremely small values in the north may be overestimated while the extremely large values in the south may be underestimated to some extent. The overestimation and underestimation issues are quite common in machine learning (Cho, Yoo, Im, & Cha, 2020; Li, Li, Li, & Liu, 2020; Uddin, Nash, Mahammad Diganta, Rahman, & Olbert, 2022; Wu et al., 2022). This can explain the temporal variability of the bias (Figure 8) as well. Positive bias values are more likely to be seen in early periods with lower temperature, and negative bias values tend to appear in more recent periods with higher temperature. Although the biases have spatial and temporal variabilities, these variations are quite small (i.e., ranging from -0.3 °C to +0.3 °C). Overall, the estimations in our study are reliable (see the evaluation results in Section 4.1).
References
Cho, D., Yoo, C., Im, J., & Cha, D. H. (2020). Comparative Assessment of Various Machine Learning‐Based Bias Correction Methods for Numerical Weather Prediction Model Forecasts of Extreme Air Temperatures in Urban Areas. Earth and Space Science, 7(4). doi:https://doi.org/10.1029/2019ea000740
Li, Y., Li, M., Li, C., & Liu, Z. (2020). Forest aboveground biomass estimation using Landsat 8 and Sentinel-1A data with machine learning algorithms. Sci Rep, 10(1), 9952. doi:https://doi.org/10.1038/s41598-020-67024-3
Uddin, M. G., Nash, S., Mahammad Diganta, M. T., Rahman, A., & Olbert, A. I. (2022). Robust machine learning algorithms for predicting coastal water quality index. J Environ Manage, 321, 115923. doi:https://doi.org/10.1016/j.jenvman.2022.115923
Wu, J., Fang, H., Qin, W., Wang, L., Song, Y., Su, X., & Zhang, Y. (2022). Constructing High-Resolution (10 km) Daily Diffuse Solar Radiation Dataset across China during 1982–2020 through Ensemble Model. Remote Sensing, 14(15). doi:https://doi.org/10.3390/rs14153695
- AC6: 'Reply on RC3', Ming Luo, 17 Nov 2022
-
AC3: 'Reply on RC3', Ming Luo, 07 Nov 2022
Peer review completion
Interactive discussion
Status: closed
-
RC1: 'Comment on essd-2022-257', Minyan Wang, 23 Aug 2022
Major comments.
The dataset is very import to study the climate change during this period over this region. The high spatial resolution and the monthly temporal resolution are basically reasonable, based on the previous datasets. The use of homogenous meteorological observation data is the right way to evaluate. This research is novel and belongs to the application of big data for large volume of Satellite data used, and machine learning algorithm. The authors are from 8 universities/colleges. The results supply important reference for other research fields like the field of geoscience, biology, sociology, and medicine and engineering.
- Line 415, harm, it is suggested to change the word. All 11 indices are calculated based on SAT, so the accuracy of this dataset is up to the accuracy of the dataset (Zhang 2022b) and the algorithm. SAT has important influence to the other 11 indices. How to understand these?
- (Zhang 2022b) SAT is LST, but not Tair (1.5 meters above the surface). Why use LST but not Tair in Table 1 to compute the other 11 indices?
- Line 157-158 “Section 6 compares our products with two existing datasets, and the main findings of this paper are summarized in Section 7”. Where is section 6 and section 7? They are data availability and conclusions.
- Why not keep Figures S1-S9, Tables S1-S5 as as formal ones? Please consider again which to keep in the manuscript.
- How to get monthly data from MODIS daily mid-daytime 13:30 and mid-nighttime 01:30 LST? The monthly mean is different considering of the diurnal variation and satellite observations from ascending orbits and descending orbits.
- Line 190-191, how to compute and use the covariates?
- HiTIC? What is i? Is it HTI (human thermal index)? No “I” after “human” and before ” thermal”.
Minor comments.
- Add abbr. of LGBM when it is first used. Light Gradient Boosting Machine. Check others, please.
- Author contribution. It seems only H.Z is responsible for analysis and data processing, others are all involved in writing. For computations, are there anyone else in the author list?
- Line 439. The unit of the dataset is degree or 0.01 degree? Is 0.01 the scale factor?
- Please revise the followings.
- Line 247-248, as the figure displays
- Line 384, Figures S8j&m. “Figure S8”, not figures, and add a blank before j.
- It is suggested to delete equations (1-2) and (3-6), add the reference.
- If necessary, please keep the same description of the time period for there are many kinds in the manuscript, like 2003~2020, from 2003 to 2020, during the year 2003~2020, (2003 to 2020), during 2003~2020, from January 2003 to December 2020, 2003-2020.
- Line 755. Delete the last sentence. Line 759, change : to . after 2003~2020.
- Add longitude and latitude (or the numerical scope signs) in Fig 1, 4-7, 9-12, 14.
- Figure 8. Prediction accuracies of 12 human thermal indices… Add 12. In individual years, change the description. Time series? Add the description of (a) (b) (c).
- Figure 10. Spatial distributions of 12 human thermal indices... Add 12.
- Line 780. Figure 11. National average, are you sure to write like this? Check it in other places in the manuscript. Just use average is fine. Delete “straight”. Line 789, Figure 13, national, delete this. Add 12, 12 human thermal indices.
- Line 785, Figure 12, add 12. “the trends of annual mean”, please consider how to express better. Inter-annual variation?
- Line 795, Figure 14. HITIC, is it “i”? in mainland China, is it over mainland of China? Please check the description of “mainland China” in the manuscript. In July 2018, move to the end of four major UAs. Delete i.e., change to :.
- AC1: 'Reply on RC1', Ming Luo, 01 Oct 2022
- AC4: 'Reply on RC1', Ming Luo, 17 Nov 2022
-
RC2: 'Comment on essd-2022-257', Anonymous Referee #2, 28 Aug 2022
The authors have produced a high-resolution (1 km×1 km) thermal index collection at a monthly scale (HiTIC-Monthly) in China during 2003 to 2020, with 12 widely used human thermal indices. The authors have created a high-resolution products for quantifying thermal index in China, which is valuable to the scientific community. I have some comments to be addressed by the authors.
1. The biggest concern is the temporal resolution. Why do the authors choose monthly resolution, rather than daily? Daily products would be extremely useful to characterize extreme events, which are of societal importance.
2. Lines 168-169: How about the impacts of precipitation on thermal indices? Have the authors considered precipitation as a covariate?
3. Figures 7 and 8: The results indicate spatial variability of bias in the thermal indices. What factors drive the spatial variability of the bias? Meanwhile, there is also temporal variability in the bias (Figure 8), and what is the drivers of this variability? Are the spatial and temporal variabilities of the bias related to background climates?
4. One way of evaluating the quality of these products is to evaluate the EOFs of these products. For example, what are the first three EOFs in each product? How do the temporal coefficients change over time across these products? Such spatial-temporal evaluation would be desirable.- AC2: 'Reply on RC2', Ming Luo, 01 Oct 2022
- AC5: 'Reply on RC2', Ming Luo, 17 Nov 2022
-
RC3: 'Comment on essd-2022-257', Minyan Wang, 26 Oct 2022
3) Line 157-158, suggest: Comparisons on our products with two existing datasets are in Section 5, data availability is provided in Section 6, ...
The 3rd response to the reviewer2 (spatial variability, temporal variability, background climates), Please consider how to answer the question directly.
-
AC3: 'Reply on RC3', Ming Luo, 07 Nov 2022
1.Line 157-158, suggest: Comparisons on our products with two existing datasets are in Section 5, data availability is provided in Section 6, ...
Response: Revised per your suggestion.
2.The 3rd response to the reviewer2 (spatial variability, temporal variability, background climates), Please consider how to answer the question directly.
Response: Thank you very much for your comment. Below please see our updated and direct responses:
As shown in Figure 7, the biases exhibit zonal variations across the space, i.e., positive bias values tend to distribute in northern China and negative values are mainly located in the south. This spatial variability is likely caused by the generally low and high temperatures in the north and south, respectively. The extremely small values in the north may be overestimated while the extremely large values in the south may be underestimated to some extent. The overestimation and underestimation issues are quite common in machine learning (Cho, Yoo, Im, & Cha, 2020; Li, Li, Li, & Liu, 2020; Uddin, Nash, Mahammad Diganta, Rahman, & Olbert, 2022; Wu et al., 2022). This can explain the temporal variability of the bias (Figure 8) as well. Positive bias values are more likely to be seen in early periods with lower temperature, and negative bias values tend to appear in more recent periods with higher temperature. Although the biases have spatial and temporal variabilities, these variations are quite small (i.e., ranging from -0.3 °C to +0.3 °C). Overall, the estimations in our study are reliable (see the evaluation results in Section 4.1).
References
Cho, D., Yoo, C., Im, J., & Cha, D. H. (2020). Comparative Assessment of Various Machine Learning‐Based Bias Correction Methods for Numerical Weather Prediction Model Forecasts of Extreme Air Temperatures in Urban Areas. Earth and Space Science, 7(4). doi:https://doi.org/10.1029/2019ea000740
Li, Y., Li, M., Li, C., & Liu, Z. (2020). Forest aboveground biomass estimation using Landsat 8 and Sentinel-1A data with machine learning algorithms. Sci Rep, 10(1), 9952. doi:https://doi.org/10.1038/s41598-020-67024-3
Uddin, M. G., Nash, S., Mahammad Diganta, M. T., Rahman, A., & Olbert, A. I. (2022). Robust machine learning algorithms for predicting coastal water quality index. J Environ Manage, 321, 115923. doi:https://doi.org/10.1016/j.jenvman.2022.115923
Wu, J., Fang, H., Qin, W., Wang, L., Song, Y., Su, X., & Zhang, Y. (2022). Constructing High-Resolution (10 km) Daily Diffuse Solar Radiation Dataset across China during 1982–2020 through Ensemble Model. Remote Sensing, 14(15). doi:https://doi.org/10.3390/rs14153695
- AC6: 'Reply on RC3', Ming Luo, 17 Nov 2022
-
AC3: 'Reply on RC3', Ming Luo, 07 Nov 2022
Peer review completion
Journal article(s) based on this preprint
Hui Zhang et al.
Data sets
HiTIC-Monthly: A High Spatial Resolution (1 km×1 km) Monthly Human Thermal Index Collection over China from 2003 to 2020 Hui Zhang, Ming Luo, Yongquan Zhao, et al. https://zenodo.org/record/6895533
Hui Zhang et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 639 | 186 | 24 | 849 | 37 | 11 | 12 |
- HTML: 639
- PDF: 186
- XML: 24
- Total: 849
- Supplement: 37
- BibTeX: 11
- EndNote: 12
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(3526 KB) - Metadata XML
-
Supplement
(2228 KB) - BibTeX
- EndNote