16. Conclusion
• The algorithms for scale transformation are applied to
downscaling and to upscaling the surfaces of elevation,
temperature, precipitation and population
• The applications demonstrated that the HASM based algorithms
have much higher accuracy comparing with other algorithms
• Further clarification and assessment of the advantages and
disadvantages of the various options for coupling HASM and
related models is another focus of future work
17. References
• Tian-Xiang Yue, et al. 2016. A fundamental theorem of Earth’s surface modelling. Environmental Earth
Sciences 75(9): article 751 (pages 1 -12).
• Zhang, B., Fan, Z., Du, Z. et al. A Geomorphological Regionalization using the Upscaled DEM: the
Beijing-Tianjin-Hebei Area, China Case Study. Sci Rep 10, 10532 (2020). https://doi.org/10.1038/s41598-
020-66993-9
• Tianxiang YUE, Na ZHAO, Yu LIU, Yifu WANG, Bin ZHANG, Zhengping DU, Zemeng FAN, Wenjiao SHI,
Chuanfa CHEN, Mingwei ZHAO, Dunjiang SONG, Shihai WANG, Yinjun SONG, Changqing YAN, Qiquan
LI, Xiaofang SUN, Lili ZHANG, Yongzhong TIAN, Wei WANG, Ying’an WANG, Shengnan MA,
Hongsheng HUANG, Yimin LU, Qing WANG, Chenliang WANG, Yuzhu WANG, Ming LU, Wei ZHOU, Yi
LIU, Xiaozhe YIN, Zong WANG, Zhengyi BAO, Miaomiao ZHAO, Yapeng ZHAO, Yimeng JIAO, Ufra
NASEER, Bin FAN, Saibo LI, Yang YANG, John P. WILSON, A fundamental theorem for eco-
environmental surface modelling and its applications, In Journal of SCIENCE CHINA Earth Sciences,
Volume 63, Issue 8, 2020, Pages 1092-1112, ISSN 1674-7313, https://doi.org/10.1007/s11430-019-9594-
3.
• Zhao, N., Chen, C., Zhou, X. et al. A comparison of two downscaling methods for precipitation in
China. Environ Earth Sci 74, 6563–6569 (2015). https://doi.org/10.1007/s12665-015-4750-7
• Eco-Environmental Informatics ,2020.10.19, Section 3.
A method for high accuracy surface modelling (HASM) has been developed since 1986. we are well known about that HASM produced more accurate results than the classical methods. So HASM based upscaling and downscaling has been established to acquire both extrinsic information and intrinsic information from the surface.
Hello ladies and gentlemen , I am suraj shah from College of resources and environment ,UCAS and the topic of my todays presentation is HASM-based upscaling(HASM-US) and HASM-based downscaling(HASM-DS)
without any delay ,lets jump into outlines of my presentation….
Well, after short description of my topic for today presentation, lets jump into outlines ,In this presentation I am going to introduce the notion of four topics..
at first about General concepts about upscaling and downscaling
Then at second I will introduce simple definition with core algorithm of both methodologies..
then at 3rd I will show comparison of these HASM based methodologies with other classic concepts and some well established papers conclusion to support advantages of these newly well established methods
then at last I will wrap my presentation with conclusion
With these information I would like to move on general concepts…which is on next slide..
Information from earth surface are represented by Macro patterns and micro processes with provide abundant streams of information
we have as scenario when ,spatial heterogeneity on the micro-scale may not be detected at a coarse spatial resolution, and conversely, general patterns on the macro-scale may not be detected at a fine spatial resolution
Thus there is a need of method of upscaling and downscaling .
Where
Upscaling refers to the transfer of knowledge from a finer to a coarser resolution in order to reduce computational costs
Downscaling approaches have been developed to obtain information at finer resolution from the coarser resolution models
Moreover we can see figure representing
With this description of needs and adopted methodology of upscaling and downscaling now I like to move on individual concept of HASM based Upscaling ..
Grids with different spatial resolutions within the same region, Ω(H) and Ω(h), are expressed as follows
The spatial resolutions of these two grids are expressed as H and h, where H > h
i and j denote the number of rows and columns in the grid
where AH denotes the coefficient matrix of grid Ω(H), XH +(n 1) is the unknown parameter, bH (n) is the right item, CH is the coefficient matrix of the optimum control equation, and dH (n) is the right end item of the optimal control equation.
A simple arithmetic averaging method was then adopted to reduce computation, and a value of XH(I, J) was obtained for more accurate simplification using the inverse distance interpolation method. The upscaling process of each coarse grid can therefore be similarly simplified; this calculation process can be defined as a grid projection operator GhH,
As XH is the trend surface, it is possible to calculate the residual of the sample point set П and the trend surface, which enables HASM simulation to be performed on residuals. A grid residual surface was therefore obtained by superimposing the trend surface, XH(I, J), to obtain the final DTM. This upscaling algorithm was defined as HASM-US and was formulated using equation as shown …
Upscaling elevation surface on spatial resolution of 30m×30m to 1000m×1000m
Applying the HASM algorithm, a HASM-upscaling (HASM-US) algorithm for DEM upscaling simulation was
developed in this study
First, cross validation and surface profile analyses reveal that HASM-US enables higher accuracy and better reflects extreme values within DEMs in terms of upscaling simulation results when compared to those of either bilinear interpolation as margin of error was in the difference of 230 m
. HASM-US is therefore a more efficient method for upscaling DEMs
1.A fundamental theorem for eco-environmental surface modelling and its applications applied on The Beijing-Tianjin-Hebei (BTH) region is taken as a case area to conduct empirical studies of algorithms for spatial upscaling,
The cross-validation results indicated that the MPAE,MNAE and RMSE of HASM-US are 59, −70 and 19 m, respectively, whereas those for BR are 72, −107 and 28 m and those for ARP are 77, −113 and 28 m, respectively
The spatial resolutions of many models or data are sometimes too coarse to be used for analyses at regional or local scales. To overcome this problem, many downscaling approaches have been developed to obtain information at finer resolution from the coarser resolution models and data
it is the case of downscaling of elevation surface in which HASM based Down scaling achieved minimum error of 0.84%
The validation results show that HASM-DS has much higher accuracy than the traditional downscaling method, ordinary linear squares (OLS-DS)
Mean absolute errors (MAEs) of the original MAT from CMIP5, downscaled MAT by OLS-DS and downscaled MAT by HASM-DS are 1.86, 1.62 and 0.51°C, respectively; their mean relative errors are 24%, 23% and 7% and their correlation coefficients are 0.66, 0.80 and 0.96, respectively.
In short, the accuracy of the original MAT was improved by 17% using HASM-DS (p=0.05).
It is the case study of , downscaling elevation surfaces
downscaling CMIP5 scenarios from 100km×100km to 1km ×1km
downscaling population statistics from provincial level to 1km×1km level
Downscaling population statistics from township level to 90m×90m level