An interpolated temperature and precipitation dataset at 1-km grid resolution in China (2000 – 2012)

No comments yet

Submit questions or advice:



You are not logged in, please[Login]or[ Register]!

An interpolated temperature and precipitation dataset at 1-km grid resolution in China (2000 – 2012)

Total number of views and downloads

View in HTML Paper download
3069 50

An interpolated temperature and precipitation dataset at 1-km grid resolution in China (2000 – 2012)

The author's papers

Sorry, failed to retrieve the author's related papers.

            Data source: Chinese Science Citation Database(CSCD)

An interpolated temperature and precipitation dataset at 1-km grid resolution in China (2000 – 2012)

Wang Junbang1*, Wang Juwu2, Ye Hui1, Liu Ya1, He Honglin1

1. Key Laboratory of Ecosystem Network Observation and Modeling, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, P. R. China;

2. College of Resources and Environment, Northeast Agricultural University, Harbin 150030, P. R. China

*Email: jbwang@igsnrr.ac.cn

Abstract: With the rapid development of satellite remote sensing techniques with high spatial resolution and their wide applications in the field of ecology, high-resolution meteorological spatial data have become popular in ecosystem process simulation systems, especially on spatiotemporal dynamics of terrestrial ecosystems with high spatial heterogeneity in the background of global climate change. Spatial meteorological data with high resolution consistent with the land surface data from MODIS is essential to estimating the net primary productivity of vegetation in terrestrial ecosystems of China. The temperature and precipitation datasets with a grid resolution of 1 km were produced based on the observed data from the National Meteorological Information Center (NMIC) of China Meteorological Administration and the Daily Global Historical Climatology Network-Daily (GHCN-D) during 2000–2012. The stations were selected in a region larger than the research area and data quality was evaluated. The eight-day average and total values were calculated for temperature and precipitation, respectively. Finally, the ANUSPLIN software was used to interpolate the observed meteorological data to a grid with high resolution (1-km). Those two annual data products can be used to further study the spatiotemporal changes in terrestrial ecosystems of China.

Keywords: temperature; precipitation; interpolation; 1 km grid; China

Dataset Profile

Chinese title

2000~2012年全国气温和降水1 km网格空间插值数据集

English title

An interpolated temperature and precipitation dataset at 1-km grid resolution in China (2000 – 2012)

Corresponding author

Wang Junbang (jbwang@igsnrr.ac.cn)

Data authors

Wang Junbang, Wang Juwu, Ye Hui, Liu Ya, He Honglin

Time range

2000–2012

Geographic scope

Mainland China

Spatial resolution

1-km grid

Data volume

5.1 GB

Data format

32-bit floating type in ArcGIS format (*.flt)

Data service system

<http://www.cnern.org.cn/data/initDRsearch?classcode=DPAPER>

<http://www.sciencedb.cn/dataSet/handle/319>

Sources of funding

Science and Technology Service Network Initiative of Chinese Academy of Sciences (STS Plan, KFJ-SW-STS-167);

National Key Technologies R&D Program (2016YFC0500203);

National Natural Science Foundation of China (31270520)

Dataset composition

The dataset is composed of interpolated precipitation data (2.58 GB) and interpolated temperature data (2.51 GB) (2000–2012). File names are PRCP and TEM, respectively.

1. Introduction

Meteorological data are the foundation of research on long-term changes in terrestrial ecosystems. However, with the advances in remote sensing techniques, the ecosystem research is mostly conducted by using spatial data with resolutions ranging from 30 m to 1 km instead of focusing on single points at sampling sites or the patch scale1-3. Meteorological data must match the temporal and spatial resolution of the remote sensing data to study long-term, dynamic changes in spatial patterns. Current meteorological data based on global climatology are not conducive to the analysis of regional climate change; therefore, it is crucial to collect meteorological data with high spatiotemporal resolution4-6.

The daily surface climate dataset for 2000–2012 established by the National Meteorological Information Center (NMIC) of China Meteorological Administration and the Daily Global Historical Climatology Network-Daily (GHCN-D) contains daily total precipitation and mean temperature data at a grid resolution of 1 km generated by spatial interpolation methods applying two software programs: the MATLAB-Based Spatial Interpolation Preprocessing System of Observation Data from Main International Meteorological Stations developed by the Key Laboratory of Ecosystem Network Observation and Modeling, Chinese Academy of Sciences (Software registration number: 2014SR142569 and software classification number: 30200-7500) and widely used software, ANUSPLIN, developed by Australian National University. The data in these datasets have consistent spatiotemporal resolutions with the data products for the land surface collected by the MODerate-resolution Imaging Spectroradiometer (MODIS) instruments. The dataset provides fundamental meteorological data for the estimation of the net primary productivity of vegetation and data support for the study of spatiotemporal changes in terrestrial ecosystems of China.

Figure 1 Distribution of digital elevation model (DEM) and meteorological stations

2. Data collection and processing

The dataset is composed of data from two sources: GHCN-D and the daily surface climate dataset established by the NMIC of China Meteorological Administration, including daily precipitation and mean temperature data collected from 345 and 753 ground meteorological stations, respectively (Figure 1). Those data were standardized using the MATLAB-Based Spatial Interpolation Preprocessing System of Observation Data from Main International Meteorological Stations. DEM was built using elevation data at 90-m resolution from the International Scientific Data Service Platform of Chinese Academy of Sciences. DEM data were then converted to ASCII data files that can be read by ANUSPLIN (Figure 2).

Figure 2 Data preprocessing and interpolation

2.1 Interpolation algorithm in ANUSPLIN

The spatial interpolation algorithm is the core of producing meteorological grid data. ANUSPLIN is a tool for conversion analysis and interpolation of multivariate data and uses the thin plate smoothing splines algorithm, commonly used in Australia, the US, and Europe7-10. This software provides a means for comprehensive statistical analysis, data diagnostics, and spatially-distributed standard error analysis. It also supports flexible data input and fitted surface output. Among the eight ANUSPLIN programming modules, only the main ones should be executed, including SPLINA (thin plate smoothing spline function) and LAPGRD (used to calculate values and Bayesian standard error estimates of partial thin plate smoothing spline surfaces), since data were collected at less than 2,000 stations. The core algorithm in ANUSPLIN is the partial thin plate smoothing spline method, with the following theoretical statistics model11:

where zi is the dependent variable at point i; f (xi) is the unknown smoothing function of xi; xi is the independent variable of the d-dimensional spline i; bT is the p-dimensional coefficient of yi; yi is the independent covariant of p-dimension; ei is the random error of the independent variable with the expectation value of 0 and the variance of ; wi is the known relative error variance; is the constant of error variance of all unknown stations; and N is the number of observed values. When the formula (1) lacks the second part, i.e., there is no covariant (p = 0), the model can be simplified to the general thin plate smoothing spline model. When the formula (1) lacks the first part, the model will become the simple multiple linear regression model, but ANUSPLIN does not allow the occurrence of the second condition.

Function f and coefficient b are determined using the following least squares estimation:

where Jm (f) is the coarseness measure as a function of f, i.e. m-order partial derivative of function f. \\rho is the smoothing parameter of positive value. When \\rho→0, the function f is close to the accurate interpolation formula; when \\rho→+∞, the function f is close to a least squares polynomial, depending on coarseness m. The value of the smoothing parameter is usually dependent on the minimum prediction error of the fitted surface and is determined by generalized cross validation (GCV). GCV is performed by excluding data points one by one and calculating the residual error of a specific point using other data points under the condition of the same \\rho. The calculation records are provided in the output log files of ANUSPLIN.

2.2 Data processing

Data processing included the steps of inputting data, dataset merging, data evaluation, data output, generation of scripts for batch processing, and data interpolation using ANUSPLIN. The input data mainly included the geographical information of the meteorological stations, DEM data, GHCN-D data, and the daily surface climate data. After input data were read and merged, the format was standardized and data completeness was examined. The stations with more than 10% missing data points were labeled as invalid and excluded from the analysis. The missing records of valid stations were filled using the inverse distance weighted interpolation method to ensure the reliability of the interpolation process and results and overcome the problem of uneven distribution of stations. The station data of every eight days were averaged for temperature and summed for precipitation. The annual mean temperature and annual total precipitation were calculated accordingly. The main data sources included DEM, daily surface climate data, and GHCN-D data.

(1) DEM data

DEM data are in ASCII format. They were downloaded from the International Scientific Data Service Platform of Chinese Academy of Sciences and generated from the NASA’s Shuttle Radar Topography Mission (SRTM) GRID data. Part of these data was used depending on the scope of the study.

(2) Chinese daily surface climate data

These data were derived from the Monthly Reports on Surface Meteorology in provinces, cities, and autonomous regions.

(3) GHCN-D data

The data were downloaded from the website <ftp://ftp.ncdc.noaa.gov/ pub/data/ghcn/ daily>.

3. Sample description

The final dataset was named as PRCP_YYYY.flt or TEM_YYYY.flt (e.g., PRCP_2006.flt and TEM_2006.flt). The file name presents information about the dataset. PRCP represents the precipitation grid data with a 1-km spatial resolution; TEM refers to the temperature grid data with a 1-km spatial resolution; and YYYY is the year. In Figure 3, a and b show the annual total precipitation in 2002 and 2006, respectively; c and d illustrate the annual mean temperature in 2002 and 2006, respectively.

Figure 3 Spatial patterns of precipitation and temperature at 1-km grid resolution based on the observed data collected from the meteorological stations

4. Quality control and assessment

To verify the precision of the dataset, we adopted the data of the key meteorological factors from seven flux monitoring towers (http://asiaflux.net) in Asian areas (Figures 4, a and c are the scatter diagrams of the spatial interpolation data and the observed data from carbon flux monitoring towers and show the fitting results; Figures b and d illustrate the frequency distribution). This method overcame the shortcomings of the traditional cross-checking methods, ensuring data reliability. The interpolation results were significantly linearly correlated with the observed data from the flux monitoring towers. The multiple correlation coefficient (R2) of the linear regression was 0.67 for precipitation (p < 0.01) and the root-mean-square error (RMSE) was 18.50 mm (Figure 4a). The frequency of the observed and interpolated precipitation data showed better consistency (Figure 4b). The performance of the interpolated temperature data was satisfactory with an R2 of 0.94 (p < 0.01), suggesting that the interpolated results could explain the spatial and temporal variability of 94% of the observed temperature data from the towers (Figure 4c).

Figure 5 shows the comparison between the 0.5°×0.5° grid data of surface temperature and precipitation (V2.0) and the corresponding interpolated data used in this study. Except for the obvious differences in precipitation in eastern coastal areas, the overall regional distributions of precipitation and temperature data were consistent.

Figure 4 Comparison of interpolated precipitation and temperature data collected from the meteorological stations and the data obtained from the carbon flux monitoring towers

Figure 5 Comparison between the two interpolated datasets of precipitation and temperature in China

The subplots are the interpolated precipitation (a, b) and temperature (c, d) of 1-km grid size in this study (a, c) and the 0.5°×0.5° grid size produced by the China Meteorological Administration (b, d, geographic scope of this dataset was Mainland China, so data were missing).

5. Usages notes

The interpolated dataset of temperature and precipitation from 2000 to 2012 can be downloaded from the data resources service website (http://www.cnern.org.cn) of the Chinese Ecosystem Research Network (CERN) or Science Data Bank (http://www.sciencedb.cn/dataSet/handle/319) after clicking “Data for Papers” The dataset contains the binary grid data of floating-point type output by ArcGIS, i.e., in ArcGIS float format. Except for the 32-bit floating data file with a file extension of .flt, there are also a header file and a projection information file with the same filename with extensions of .hdr and .prj, respectively. These files could be read and re-analyzed by ArcGIS or MATLAB.

Acknowledgements

We extend our sincere gratitude to the National Meteorological Information Center of China Meteorological Administration who provided the original data as well as the Chinese Ecosystem Research Network and the Scientific Data Center of the Computer Network Information Center of Chinese Academy of Sciences who provided a data storage platform and technical support.

References

1. Chapin III F, Matson P, Mooney H. Principles of terrestrial ecosystem ecology. New York: Springer Science & Business Media, 2011: 369–397.

2. Bellard C, Bertelsmeier C, Leadley P et al. Impacts of climate change on the future of biodiversity. Ecology Letters 15 (2012): 365–377.

3. Frey S, Lee J, Melillo J et al. The temperature response of soil microbial efficiency and its feedback to climate. Nature Climate Change 3 (2013): 395–398.

4. Schuur E, Mcguire A, Schädel C et al. Climate change and the permafrost carbon feedback. Nature 520 (2015): 171–179.

5. Hulme M. A 1951–80 global land precipitation climatology for the evaluation of general circulation models. Climate Dynamics 7 (1992): 57–72.

6. Fan Y, Dool H. A global monthly land surface air temperature analysis for 1948–present. Journal of Geophysical Research: Atmospheres 113(2008).

7. Hutchinson M. The application of thin plate smoothing splines to continent-wide data assimilation. BMRC research report 27 (1991): 104–113.

8. Hutchinson M. Interpolation of rainfall data with thin plate smoothing splines. Part I: Two dimensional smoothing of data with short range correlation. Journal of Geographic Information and Decision Analysis 2 (1998): 139–151.

9. Hutchinson M. Interpolation of rainfall data with thin plate smoothing splines. Part II: Analysis of topographic dependence. Journal of Geographic Information and Decision Analysis 2 (1998): 152–167.

10. Hutchinson M. ANUSPLIN Version 4.2 User Guide. Canberra: Center for Resourceand Environmental Studies, the Australian National University, 2001.

11. Shang Z, Gao Q, Yang D. Spatial pattern analysis of annual precipitation with Climate Information System of China. Journal of Ecology 21 (2001): 689–694.

Data citation

1. Wang J, Wang J, Ye H et al. An interpolated temperature and precipitation dataset at 1-km grid resolution in China (2000 – 2012). Science Data Bank. DOI: 10.11922/sciencedb.319

Authors and contributions

Wang Junbang, PhD, Associate Professor; research area: ecosystem ecology. Contribution: overall scheme design.

Wang Juwu, MSc candidate; research area: land usage and planning. Contribution: basic data collection and manuscript writing.

Ye Hui, PhD candidate; research area: carbon cycle model and ecological informatics. Contribution: basic data processing and spatial interpolation of rainfall and temperature data at 1-km grid resolution.

Ya Liu, MSc candidate; research area: land ecology and synthetic application of 3S. Contribution: basic data collection.

He Honglin, PhD, Professor; research area: ecological informatics. Contribution: overall scheme design.

 

How to cite this article: Wang J, Wang J, Ye H et al. An interpolated temperature and precipitation dataset at 1-km grid resolution in China (2000 – 2012). China Scientific Data 1 (2017). DOI: 10.11922/csdata.170.2016.0112

Download