Full-coverage 250 m monthly aerosol optical depth dataset (2000–2019) amended with environmental covariates by an ensemble machine learning model over arid and semi-arid areas, NW China

[摘要] Aerosols are complex compounds that greatly affect the global radiationbalance and climate system and even human health; in addition, aerosols arecurrently a large source of uncertainty in the numerical simulation process.The arid and semi-arid areas have fragile ecosystems with abundant dust butlack related high-accuracy aerosol data. To solve these problems, we use thebagging trees ensemble model, based on 1 km aerosol optical depth (AOD) dataand multiple environmental covariates, to produce a monthlyadvanced-performance, full-coverage, and high-resolution (250 m) AOD product(named FEC AOD, fusing environmental covariates AOD) covering the arid andsemi-arid areas. Then, based on the FEC AOD products, we analyzed thespatiotemporal AOD pattern and further discussed the interpretation ofenvironmental covariates to AOD. The results show that the bagging treesensemble model has a good performance, with its verification R 2 values always remaining at 0.90 and the R 2 being 0.79 for FEC AODcompared with AERONET AOD product. The high-AOD areas are located in theTaklimakan Desert and on the Loess Plateau, and the low-AOD areas areconcentrated in southern Qinghai province. The higher the AOD, thestronger the interannual variability. Interestingly, the AOD reflected adramatic decrease on the Loess Plateau and an evident increase in thesouth-eastern Taklimakan Desert, while the southern Qinghai province AODsshowed almost no significant change between 2000 and 2019. The annualvariation characteristics show that the AOD was largest in spring ( 0.267±0.200 ) and smallest in autumn ( 0.147±0.089 ); the annual AODvariation pattern showed different features, with two peaks in March andAugust over Gansu province but only one peak in April in otherprovinces/autonomous regions. Farmlands and construction lands have high AODlevels compared to other land cover types. Meteorological factorsdemonstrate the maximum interpretation ability of the AODs on all settemporal scales, followed by the terrain factors, while surface propertieshave the smallest explanatory abilities; the corresponding averagecontributions are 77.1 %, 59.1 %, and 50.4 %, respectively. Thecapability of the environmental covariates to explain the AOD variesseasonally in the following sequence: winter (86.6 %) > autumn(80.8 %) > spring (79.9 %) > summer (72.5 %). Inthis research, we provide a pathbreaking high spatial resolution (250 m)and long time series (2000–2019) FEC AOD dataset covering arid and semi-aridregions to support atmospheric and related studies in northwest China; thefull dataset is available at https://doi.org/10.5281/zenodo.5727119 (Chen etal., 2021b).

[学科分类] 眼科学