关键词: 电力消费量, 空间化, DMSP/OLS, NDVI, 人居指数

Abstract:

The satellite-measured DMSP/OLS nighttime light data was widely used for regional level mapping of socioeconomic activities due to its high temporal resolution, free availability and wide swath. However, the use of DMSP/OLS nighttime light data as covariates for mapping socioeconomic activities faces numbers of problems. One of these is the spatial resolution of the available data. Although the DMSP/OLS sensor has a nominal resolution of 1 km, this has been resampled from the 2.7 km native resolution of the sensor. The second difficulty is caused by “overglow” due to surface reflection and scattering and refraction in the atmosphere which results in the overestimation of lighted areas. The third problem relates to low radiometric resolution of 6 bits (i. e. the digital number value ranges from 0 and 63) which results in data saturation over brightly light built-up areas. Vegetation indexes like NDVI are negatively correlated with the impervious surfaces and can be used for estimation of built-up areas. The incorporation of NDVI can reduce the errors occurring in estimating built-up areas from the DMSP/OLS nighttime light imagery due to data saturation and other factors. In present study, the DMSP/OLS nighttime light data was combined with SPOT NDVI data to develop an index called human settlement index (HSI), which estimated the fraction of built-up area on a per pixel basis. Due to the complementary characteristics between DMSP/OLS data and NDVI, the resultant HSI image conveys more information than both the individual datasets. The model for electricity consumption estimation was developed based on the significant correlation between the HSI and electricity consumption in Zhejiang Province in the article. Preliminary modeling results show general overestimation of electricity consumption, especially in high altitude area in southwest Zhejiang Province. The HSI was further corrected by thresholding method to overcome the overglow effect and elevation effect correction was also conducted. The modified HSI image was then used for mapping the electricity consumption in 2010 in Zhejiang Province at a resolution of 1 km×1 km. The results show that the correction of HSI results in a significant increase in accuracy in mapping the electricity consumption. The mean relative error is 26% when modified HSI was used to estimated the electricity consumption of Zhejiang province, which is much smaller than previous studies. The spatial distribution of electricity consumption is well in line with the economic development level. In addition, more than 75% of the electricity consumption located in area below 50 m in Zhejiang Province. The present research provides an integrated approach for rapid and accurate estimation of electricity consumption in regional scale on a per pixel-basis, which can be very useful for mapping socioeconomic activities from medium coarse resolution data at regional level within limited time and minimal cost.

Key words: electricity consumption, spatialization, DMSP/OLS, NDVI, human settlement index