EN中文

地理研究 >

2008 , Vol. 27 >Issue 6: 1299 - 1307

DOI: https://doi.org/10.11821/yj2008060009

地表过程研究

中国土壤温度的空间插值方法比较

展开
  • 1. 中国科学院南京土壤研究所 土壤与农业可持续发展国家重点实验室,南京 210008;
    2. 北京市农林科学院 农业综合发展研究所, 北京 100097;
    3. 中国科学院科技政策与管理科学研究所, 北京 100190;
    4. 中国科学院研究生院, 北京 100039
张慧智(1976-), 女, 湖南常德人, 博士。主要从事土壤资源和生态环境研究。 E-mail:zhanghuizhi07@yahoo.cn

收稿日期: 2007-10-29

  修回日期: 2008-04-29

  网络出版日期: 2008-11-25

基金资助

国家自然科学基金创新群体项目(40621001);国家重点基础研究发展计划项目(2007CB407206)

收起

Spatial prediction of soil temperatures in China using different methods

Expand
  • 1. State Key Laboratory of Soil and Sustainable Agriculture, Institute of Soil Science, CAS, Nanjing 210008, China;
    2. Institute of Integrated Development of Agriculture, Beijing Academy of Agriculture and Forestry Sciences, Beijing 100097, China;
    3. Institute of Policy and Management, CAS, Beijing 100190, China;
    4. Graduate University of Chinese Academy of Sciences, Beijing 100039, China

Received date: 2007-10-29

  Revised date: 2008-04-29

  Online published: 2008-11-25

Supported by

国家自然科学基金创新群体项目(40621001);国家重点基础研究发展计划项目(2007CB407206)

Fold

摘要

利用中国698个气象站点1971~2000年的地面气候资料,采用三种不同方法预测中国0cm、20cm和40cm深度年均土壤温度的空间分布,其中普通克里格和泛克里格法直接以年均土壤温度数据为源数据、回归克里格法以中国年均气温数据和中国DEM数据为源数据进行预测。预测结果的准确性通过平均绝对误差(MAE)和均方根误差(RMSE)值来评价。结果表明回归克里格法预测的MAE值和RMSE值均为最小,说明其预测结果的准确性最好、预测的极端误差也最小;其次为泛克里格法;普通克里格法预测的效果最差。回归克里格法预测结果由于采用了中国DEM数据进行修正,在空间特征表达方面能够更好地表达复杂地形地区的局部变异,其平滑效应明显小于泛克里格法和普通克里格法的预测结果。

关键词： 土壤温度; 空间分异; 普通克里格; 泛克里格; 回归克里格

本文引用格式

张慧智, 史学正, 于东升, 王洪杰, 孙维侠, 黄宝荣 . 中国土壤温度的空间插值方法比较[J]. 地理研究, 2008 , 27(6) : 1299 -1307 . DOI: 10.11821/yj2008060009

Abstract

Soil temperature influences many biological and geochemical processes in soil. With the growing interest in the fields of environmental sciences, soil science and global change studies, the spatial data of soil temperature in raster form at a national scale are needed. However, there are only discrete data of soil temperatures in China, which are extracted from meteorological stations. These discrete data of soil temperatures cannot meet the needs now. Thus, it is really meaningful to derive spatially continuous data of soil temperature from discrete data using spatial interpolation methods. Based on the meteorological data from 698 stations in China in the period from 1971 to 2000 and the digital elevation models of China, this paper attempted to predict the spatial patterns of mean annual soil temperatures in China using three different methods. The data of mean annual soil temperatures from meteorological stations were used directly by ordinary kriging and universal kriging methods in the estimation. And the data of mean annual air temperatures and the DEM data of China were used by regression kriging method to predict the spatial patterns of mean annual soil temperatures in China. Prediction was validated using mean absolute error and root mean square error. The results from exploratory analysis on the original soil and air temperature data in China revealed that the two datasets were approximated by the normal distribution, which suggested that the two datasets could be analyzed by geostatistical methods. These three geostatistial methods were compared in the precision and bias of the estimations. The values of mean absolute error and root mean square error produced by regression kriging method were the lowest, which were 1.17-1.25℃ and 1.46-1.58℃, followed by universal kriging method and ordinary kriging method. It indicated that regression kriging method yielded more accurate results than the other two methods. The prediction by regression kriging had less smoothing effect and more details of depicting the local variation under complex terrain. Therefore, regression kriging was the best method for predicting mean annual soil temperatures in China in this study. This was because the regression residuals were incorporated within kriging system in the regression kriging method, and the DEM data were used as auxiliary information to improve the accuracy of predictions of mean annual soil temperatures.

Key words： soil temperature; spatial variability; ordinary kriging; universal kriging; regression kriging

参考文献


[1] Fitton E M, Brooks C F. Soil temperatures in the United States. Monthly Weather Review, January, 1931:6~16.

[2] Paul K I, Polglase P J, Smethurst P J, et al. Soil temperature under forests: A simple model for predicting soil temperature under a range of forest types. Agricultural and Forest Meteorology, 2004, 121:167~182.

[3] Plauborg F. Simple model for 10 cm soil temperature in different soils with short grass. European Journal of Agronomy,2002, 17:173~179.

[4] Gutman G. Contribution of the NASA and-Cover/Land-Use Change Program to the Northern Eurasia Earth Science Partnership Initiative: An overview. Global and Planetary Change, 2007, 56(3-4): 215~234.

[5] 王英, 曹明奎,陶波,等.全球气候变化背景下中国降水量空间格局的变化特征.地理研究,2006,25(6):1031~1042.

[6] Toy T J, Kuhaida A J, Munson B E. The prediction of mean monthly soil temperature from mean monthly air temperature. Soil Science, 1978, 126(3): 181~189.

[7] Meikle R W, Treadway T R. A mathematical method for estimating soil temperatures. Soil Science, 1979, 128(4): 226~242.

[8] Meikle R W, Treadway T R. A mathematical method for estimating soil temperatures in Canada. Soil Science, 1981, 131(5): 320~326.

[9] 杨学明. 土壤水热状况与土壤系统分类. 土壤,1988,21(2):110~113.

[10] 史学正,邓西海.我国土壤温度状况.中国土壤系统分类进展.北京: 科学出版社, 1993.353~360.

[11] 冯学民,蔡德利. 土壤温度与气温及纬度和海拔关系的研究. 土壤学报,2004,41(3): 489~491.

[12] Barringer James R F. Meso-scale mapping of soil temperatures in the Mackenzie Basin, New Zealand. In: Proceedings of 2nd Annual Conference on Geo-Computation. August 26-29, 1997, University of Otago, Dunedin, New Zealand, 1997. 393~396.

[13] Barringer James R F, Lilburne L R. Developing fundamental data layers to support environmental modeling in New Zealand. 4th international conference on integrating GIS and environmental modeling, Banff, Alberta, Canada, 2000.1~9.

[14] 廖顺宝,李泽辉,游松财. 气温数据栅格化的方法及其比较. 资源科学, 2003, 25 (6): 83~88.

[15] Trianfilis J, Odeh I O A, McBratney A B. Five geostatistical models to predict soil salinity from electromagnetic introduction data across irrigated cotton. Soil Sci. Soc. Am. J., 2001, 65: 869~878.

[16] Zhang X Y, Sui Y Y, Zhang X D, et al. Spatial variability of nutrient properties in black soil of Northeast China.Pedosphere, 2007, 17(1): 19~29.

[17] 王红,宫鹏,刘高焕.黄河三角洲多尺度土壤盐分的空间分异.地理研究,2006, 25(4): 649~658.

[18] Baxter S J, Oliver M A. The spatial prediction of soil mineral N and potentially available N using elevation. Geoderma, 2005, 128: 325~339.

[19] Hengl T, Heuvelink G B M, Stein A. A generic framework for spatial prediction of soil variables based on regression-kriging. Geoderma, 2004, 120: 75~93.

[20] Willmott C J, Robenson S M. Climatologically aided interpolation of terrestrial air temperature. Int. J. Climatol., 1995, 15:221~229.

[21] Zhao Y C, Shi X Z, Yu D S, et al. Soil organic carbon density in Hebei Province, China: Estimates and uncertainty. Pedosphere, 2005, 15 (3): 293~300.

[22] Collins F C, Bolstad P V. A comparison of spatial interpolation techniques in temperature estimation. Third International Conference/Workshop on Integrating GIS and Environmental Modeling. 1996. http ://www.ncgia.ucsb. edu/conf/SANTA_FE_CD-ROM/sf_papers/Collins_fred/collins.html

Options
文章导航

/

〈 〉