An evaluation of satellite dust-detection algorithms in the Middle East region

In the last 15 years, the frequency, spatial extent, and intensity of dust storms have increased and it is one of the main continuously occurring environmental hazard in the Middle East region. Since dust storms generally cover a large spatial extent and are highly dynamic, satellite Earth Observation (EO) is a key tool for detecting their occurrence, identifying their origin, and monitoring their transport and state. A variety of satellite dust detection algorithms have been developed to identify dust emissions sources and dust plumes once entrained in the atmosphere. This paper evaluates the performance of five widely applied dust detection algorithms: the Brightness Temperature Difference (BTD), D-parameter, Normalized Difference Dust Index (NDDI), Thermal-Infrared Dust Index (TDI) and the Middle East Dust Index (MEDI). These algorithms are applied to Moderate Resolution Imaging Spectroradiometer (MODIS) data to detect dust-contaminated pixels during three significant dust events in 2007 in the Middle East region that originated from sources in Iraq, Syria and Saudi Arabia. The results indicate that all methods have a comparable performance in detecting dust-contaminated pixels during the three dust events with an average detection rate (between all algorithms) of 85%. However, substantial differences exist in their ability to distinguish dust from clouds and the land surface, which resulted in large errors of commission. Direct validation of these algorithms with observations from seven Aerosol Robotic Network (AERONET) stations in the region found an average false detection rate (between all algorithms) of 89.6%. Although the algorithms performed well in detecting the dust-contaminated pixels their high false detection rate means it is challenging to apply these algorithms in operational context.     

Comparison and evaluation of dust detection algorithms using MODIS Aqua/Terra Level 1B data and MODIS/OMI dust products in the Middle East

Dust storms have a major impact on air quality, economic loss, and human health over large regions of the Middle East. Because of the broad extent of dust storms and also political–security issues in this region, satellite data are an important source of dust detection and mapping. The aim of this study was to compare and evaluate the performance of five main dust detection algorithms, including Ackerman, Miller, normalized difference dust index (NDDI), Roskovensky and Liou, and thermal-infrared dust index (TDI), using MODIS Level 1B and also MODIS Deep Blue AOD and OMI AI products in two dust events originating from Iraq and Saudi Arabia. Overall, results showed that the performance of the algorithms varied from event to event and it was not possible to use the published dust/no-dust thresholds for the algorithms tested in the study area. The MODIS AOD and OMI AI products were very effective for initial dust detection and the AOD and AI images correlated highly with the dust images at provincial scale (p-value <0.001), but the application of these products was limited at local scale due to their poor spatial resolution. Results also indicated that algorithms based on MODIS thermal infrared (TIR) bands or a combination of TIR and reflectance bands were better indicators of dust than reflectance-based ones. Among the TIR- based algorithms, TDI performed the best over water surfaces and dust sources, and accounted for approximately 93% and 90% of variations in the AOD and OMI AI data.     

Effect of major dust storm on optical,physical, and radiative properties of aerosols over coastal and urban environments in Western India

The present study deals with a major summertime dust storm event that occurred over the Arabian Desert, Saudi Arabia (SA), and the United Arab Emirates (UAE) and spread to Pakistan, the Arabian Sea, and Western India (in particular over Alibaug and Pune), using multi-satellite- and ground-based measurements. Analysis of the aerosol parameters retrieved over Alibaug and Pune from the ground-based MICROTOPS-II Sunphotometer, seven Aeronet Robotic Network (AERONET) stations, including Pune, and satellite (Moderate-resolution Imaging Spectroradiometer (MODIS)) and Ozone Monitoring Instrument ((OMI-Aura)) measurements shows significant changes on dusty days as compared with non-dusty days prior to and after the dust storm event. A large increase in aerosol optical depths (AODs) and decrease in Angstrom exponent (AE, α_440-870 nm), showing the presence of a larger fraction of coarser particles comparable to those of other intense dust outbreak episodes worldwide, have been found at all these sites during the dust storm event. Higher observed AODs are considered to be the combined effect of desert dust and burning biomass/biofuel-induced aerosols. AERONET-retrieved mean aerosol volume size distributions (AVSDs) on non-dusty and dusty days are bimodal in nature at the stations of Pune, Kanpur, Jaipur, Karachi, and Lahore, while at the stations of KAUST_Campus (SA) and Mezaira (UAE), AVSDs are monomodal. This indicates that at the first five stations, the aerosol system is a complex mixture of fine- and coarse-mode aerosols with coarse mode exerting a strong influence on the system. Dust-induced turbid conditions triggered significant extinction of 30–40%, in shortwave (SW) global solar irradiance, resulting in an increase of 57% and 74% in aerosol direct radiative forcing (ADRF) at Alibaug and Pune, respectively, causing perturbation in the radiation budget. The space-borne Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO)-retrieved aerosol vertical profiles reveal the presence of polluted dust (a mixture of fine- and coarse-mode aerosols), corroborating well with the aerosol behaviour captured by the MICROTOPS/AERONET measurements and MODIS retrievals over the study areas. The Dust Regional Atmospheric Model (BSC-DREAM8b) predictions are also found to be consistent with satellite retrievals, implying the ability of the model to monitor dust transport over the study region.     

Desert dust aerosols observed in a tropical humid city: a case study over Hong Kong

Observations from the AErosol RObotic NETwork (AERONET) sunphotometers, MODerate resolution Imaging Spectroradiometer (MODIS) satellite images, back-trajectory modelling and ‘in-situ’ PM10 measurements in Hong Kong confirmed that two dust storms on 16–17 April 2006 and 27–30 April 2009, with source areas in northwest China, affected the city. The impacts of the dust on the air quality of Hong Kong were quantified using aerosol optical properties from AERONET data and local PM10 (particle size less than 10 μm) concentrations. Combined analysis of back trajectories and the microphysical properties of the dust aerosols from AERONET inversion data suggest that the dust particulates are sometimes associated with industrial chemicals on arrival in Hong Kong. This is the first remote-sensing study to observe the presence and characteristics of Asian dust carried into the humid tropical region of south China.     

Sand and dust storm detection over desert regions in China with MODIS measurements**

A new approach is developed for quick detection of sand and dust storms (SDSs) over arid and semi-arid regions of the northwestern part of China, where the bright-reflecting source areas of Asian dust outbreaks are located. The Asian dust particles, once with proper conditions, can even transport across the Pacific Ocean and reach the USA and Canada. Remote-sensing data products of mineral dust near its source are deficient because of the radiance contributions of the bright surface. In this article, based on Moderate Resolution Imaging Spectroradiometer (MODIS) measurements, consecutive separation of dust cloud from bright underlying surface and water/ice cloud is completed by utilizing a refined cloud mask algorithm and the normalized difference dust index (NDDI). Thresholds are determined through statistical analysis of MODIS measurements over the Taklimakan and Gobi deserts. Validations with ground observations over the sites in Inner Mongolia and Xinjiang in China demonstrated good performance of the proposed method in separating SDS from bright surface and cloud.     

Validation of satellite and model aerosol optical depth and precipitable water vapour observations with AERONET data over Pune,India

ABSTRACT

The present work concerns with a detailed study of the validation of the Moderate Resolution Imaging Spectroradiometer (MODIS) and model products, and investigates the spatial and temporal variations in the correlation coefficient of the validation results obtained from the analysis of Aerosol Robotic Network (AERONET) sun–sky radiometer data archived at Pune during 2005–2015. Combining the confidence intervals and prediction levels, the ground-based AERONET aerosol optical depth (AOD) at 550 nm and precipitable water vapour (PWV) have been used to validate the MODIS, model AOD (550 nm), and PWV (cm) observations. The correlation coefficients (r) of AOD for the linear regression fits are 0.73, 0.75, and 0.79, and of PWV are 0.88, 0.89, and 0.97 for Terra, Aqua, and model simulations, respectively. Month-to-month/seasonal variation of AOD (550 nm) and PWV observations of satellite and model observations are also compared with AERONET observations. Additionally, various statistical metrics, including the root mean square error, mean absolute error, and root mean bias values were calculated using AERONET, satellite, and model simulations data. Furthermore, a frequency distribution of AOD (550 nm) and PWV observations are studied from AERONET, satellite, and model data. The study emphasizes that the globally distributed AERONET observations help to improve the satellite retrievals and model predictions to enrich our knowledge of aerosols and their impact on climate, the hydrological cycle, and air quality.     

Detection of Asia dust storms using multisensor satellite measurements

Observations from visible, infrared and microwave satellite instruments are integrated to detect dust storm over northwestern China. Microwave measurements are used to detect the dust storm underneath ice clouds, while visible and infrared measurements are utilized for delineating the cloud-free dust systems. Detection is based on microwave polarized brightness temperature differences (ΔT_b = T_bv − T_bh) among two channels of 89 GHz and 23.8 GHz and infrared brightness temperature difference (BTD) between channels at 11 and 12 μm. It is shown that the integrated approach is better than the method solely based on infrared BTD in storm detection, especially for those dust systems covered by ice clouds. This approach is applied for the Asia dust storms cases using the data from the Moderate Resolution Imaging Spectroradiometer (MODIS) and the Advanced Microwave Scanning Radiometer (AMSR-E) onboard Aqua satellite.     

Achieving downscaling of Meteosat thermal infrared imagery using artificial neural networks

This study presents the successful application of artificial neural networks (ANNs) for downscaling Meteosat Second Generation thermal infrared satellite imagery. The scope is to examine, propose, and develop an integrated methodology to improve the spatial resolution of Meteosat satellite images. The proposed approach may contribute to the development of a general methodology for monitoring and downscaling Earth’s surface characteristics and cloud systems, where there is a clear need for contiguous, accurate, and high-spatial resolution data sets (e.g. improvement of climate model input data sets, early warning systems about extreme weather phenomena, monitoring of parameters such as solar radiation fluxes, land-surface temperature, etc.). Moderate Resolution Imaging Spectroradiometer (MODIS) images are used to validate the downscaled Meteosat images. In terms of the ANNs, a multilayer perceptron (MLP) is used and the results are shown to compare favourably against a linear regression approach.     

Detection of Asian dust storms from geostationary satellite observations of the INSAT-3D imager

Asian dust storm outbreaks significantly influence air quality, weather, and climate. Therefore, it is desirable to have qualitative and quantitative information on the time, location, and coverage of these outbreaks at high spatial and temporal resolution. The imager on board the Indian metrological geostationary satellite INSAT-3D observes Asia at a temporal resolution of 30 min and a spatial resolution of 1, 4, 8, and 4 km in the visible, middle infrared (MIR), water vapour (WV), and thermal infrared (TIR) bands, respectively. In this article, an algorithm is described for detecting desert dust storms from INSAT-3D imager data. The algorithm described here is a combination of various pre-existing methods such as infrared split-window, MIR and TIR brightness temperature difference, and visible to MIR reflectance ratio, which are based on the fact that dust exhibits features of spectral dependence and contrast over the visible, MIR, and TIR spectrum that are different from clouds, surface, and clear-sky atmosphere. Using the Atmospheric Infrared Sounder (Aqua/AIRS) dust score as proxy, INSAT-3D dust storm products were tested under different scenarios such as dust storms and dust transport in Asia. TIR observations from the geostationary platform of INSAT-3D allows computation of the infrared difference dust index (IDDI), which gives a quantitative measure of dust loading relative to clear atmosphere. Moreover, due to the high temporal resolution (30 min) of INSAT-3D observations, INSAT-3D-derived dust products allow more precise monitoring of dust transportation as compared with dust products derived from polar satellite observations.     

Verification of dust forecast over the Indian region with satellite and ground based observations

This article presents the verification results of the dust forecast by a numerical model over India and neighbouring regions. National Centre for Medium Range Weather Forecasting Unified Model (NCUM) is a global numerical weather prediction (NWP) model with a prognostic dust scheme. Evaluation of the performance of dust forecast by NCUM is carried out in this study. Model forecast of dust optical depth (DOD) at 550 nm is validated against ground-based and satellite observations since optical depth measurements in mid-visible wavelength are easily available. Daily 5-day forecast based on 00 UTC initial condition during dust dominated pre-monsoon season (April–May) of 2014 is used in this study. Location specific and geographical distribution of dust forecast is validated against Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) satellite retrieved DOD observation at 532 nm, Moderate Resolution Imaging Spectroradiometer (MODIS) aerosol optical depth (AOD), Ozone Monitoring Instrument (OMI), aerosol index, and Aerosol Robotic Network (AERONET) station data of total and coarse mode AOD. The verification results indicate that NCUM dust forecast generally gives good representation of large scale geographical distribution of dust over the western region of India. DOD forecasts show good correlation with co-located CALIPSO DOD over the western part (0.71) compared to central (0.58) and eastern (0.61) part of India in April while it show similar trend in May with slightly improved correlation (0.68) over the eastern part of India. Results also show that DOD forecasts are better correlated to AERONET coarse mode AOD observations over Jaipur in April and over Kanpur in May. Vertical distribution of dust concentrations in the forecast show reasonably good agreement with attenuated backscatter and depolarization ratio from CALIPSO observations. The model is also able to simulate spatiotemporal distribution of dust during a major dust event as observed by CALIPSO, MODIS, and OMI.     

应用米氏理论选择气象卫星探测沙尘暴的波段

无论从时间尺度还是从空间尺度来看，气象卫星都是监测沙尘天气的重要手段。为了使气象卫星能在监测和预测沙尘暴中最大限度地发挥其作用，探测波段的选择是十分重要的。根据米氏理论对沙尘气溶胶的散射和消光等特性进行了计算，并根据其消光作用对气象卫星探测波段的选择作了分析。通过计算得出可见光波段是探测沙尘气溶胶的主要通道，短波红外波段对沙尘气溶胶也很敏感。同时，由于沙尘粒子对不同波段的吸收也不同，可采用不同红外探测波段的辐射亮温来综合判别沙尘区。     

Aerosol optical properties over Delhi and Manora Peak during a rare dust event in early April 2005

Dust storm events are annual phenomena observed over the Indo-Gangetic plain (IGP) during the pre-monsoon period (May–June). These dust storms affect the air quality, weather conditions and radiation budget of the region. In this paper we characterize the aerosol optical parameters associated with a rare dust storm event that hit the IGP during early April 2005. This event was considered rare as it occurred much earlier than the general occurrence of dust storms in India (May–June), and in the year 2005, the warmest year in the span of the previous hundred years.

In this study we considered the optical aerosol parameters for two places in the IGP: Delhi (28.5° N, 77.2° E, 325 m asl) and the high altitude station, Manora Peak (29.4° N, 79.5° E, 1958 m asl). Of the two selected stations, Delhi represents a highly populated and polluted location whereas Manora Peak represents a cleaner location in the central Himalayan region. During this dust storm event, the aerosol optical depth (AOD) was observed to increase considerably. The increment was 2.6–4.6 times over Delhi and 1.6–3.2 times over Manora Peak at wavelengths 380 and 1020 nm, respectively, with respect to the background values, whereas the Ångström exponent (α) for both the stations remained close to zero during the event. The effect shows a considerable increase in direct dust radiative forcing in terms of a reduction in the broadband global irradiance for Delhi as well as for Manora Peak stations. The direct aerosol radiative forcing thus obtained was about 34% in the 400–1100 nm wavelength band at Manora Peak.     

An enhanced dust index for Asian dust detection with MODIS images

An enhanced dust index (EDI) for Moderate Resolution Imaging Spectroradiometer (MODIS) solar reflectance bands is proposed that provides a means to detect the dust status of the atmosphere. The EDI utilizes only solar reflectance channels and may therefore be applied consistently to the entire MODIS time series records (1999 to present) for daytime dust observation, producing a higher spatial resolution (500 m) dust result than that from thermal-infrared records (1000 m), which were developed previously and are currently being used. The index introduces dust optical density (α), which can be simply estimated by spectral unmixing, into the normalized difference between reflectance at near-infrared (2.13 μm) and blue (0.469 μm). Dust severity can thus be rated from weak to severe within a standard range of –1 to 1. The index was applied to 11 typical dust events during 2000–2010 in East Asia, where it showed good coherence with meteorological station-observed visibility (R ² = 0.7909, p < 0.05) and standardized visibility (R ² = 0.7128, p < 0.05). Further comparison with the commonly used normalized difference dust index (NDDI) and brightness temperature difference (BTD) between MODIS bands 31 and 32 also indicated a better performance of the EDI in identifying the spatial and density distributions of dust. Previously applied satellite-based dust indices, particularly for the visible and near-infrared, can therefore be improved for a better quantification of dust aerosols.     

Nighttime polar cloud detection with MODIS

Cloud detection is the first step in studying the role of polar clouds in the global climate system with satellite data. In this paper, the cloud detection algorithm for the Moderate Resolution Imaging Spectrometer (MODIS) is evaluated with model simulations and satellite data collocated with radar/lidar observations at three Arctic and Antarctic stations. Results show that the current MODIS cloud mask algorithm performs well in polar regions during the day but does not detect more than 40% of the cloud cover over the validation sights at night. Two new cloud tests utilizing the 7.2 μm water vapor and 14.2 μm carbon dioxide bands, one new clear-sky test using the 7.2 μm band, and changes to the thresholds of several other tests are described. With the new cloud detection procedure, the misidentification of cloud as clear decreases from 44.2% to 16.3% at the two Arctic stations, and from 19.8% to 2.7% at the Antarctic station.     

The effect of soil moisture and wind speed on aerosol optical thickness retrieval in a desert environment using SEVIRI thermal channels

Dust emission and deposition are associated with several factors such as surface roughness, land cover, soil properties, soil moisture (SM), and wind speed (WS). A combination of land surface and remote-sensing models has recently been investigated for dust detection and monitoring. The thermal bands of the Meteosat Second Generation Spinning Enhanced Visible and Infrared Imager (MSG/SEVIRI) satellite are widely used for qualitative detection of dust over desert because of their high spectral and temporal resolutions. In this work, the contribution of ground-measured WS data and satellite-measured SM data on aerosol optical thickness (AOT) retrieval was investigated using an artificial neural network (ANN) model. ANNs have been applied in similar applications and have shown a higher performance than simple multiple-regression models. This performance is mainly due to the ANN's ability to capture complex and non-linear relationships between inputs and outputs. A combination of MSG/SEVIRI brightness temperature (BT)/brightness temperature differences (BTDs), BTD_3.9–10.8, BTD_8.7–10.8, BTD_10.8–12, and BT_3.9, was used as input to the base ANN model while Aerosol Robotic Network (AERONET) AOT (level 2) data at 0.5 μm were used as output. These input/output sets were obtained from two stations (Hamim and Mezaira) lying in the inland desert of the United Arab Emirates (UAE). About 3800 observations were collected, of which two-thirds were used to train the ANN model and the remaining third was kept as an independent set to assess the accuracy of the trained model. Later, Advanced Microwave Scanning Radiometer Earth Observing System (AMSR-E) SM data and ground-measured WS data were used as additional inputs to the base model to investigate their contribution to the AOT retrieval. SM data consist of daytime AMSR-E-derived daily and collected from a National Snow and Ice Data Centre (NSIDC)-archived database. Hourly average WS data were also collected at 10 m height in the same AERONET sites from two stations managed by the UAE National Centre of Meteorology and Seismology. All ground and satellite measurements were extracted for the closest time to AERONET measurements. The use of these additional inputs has been shown to have a positive impact on the accuracy of simulated AOT. The addition of these inputs to the base ANN increased R ² from 0.68 to 0.76 and reduced root mean square error from 0.113 to 0.09.     

春季沙尘暴强度的NOAA-14气象卫星监测研究

以2000~2001年发生在我国西部的春季沙尘暴为例,利用NOAA-14卫星AVHRR资料,对西北80个气象站点所对应的卫星数据进行了5个通道的信息提取。根据沙尘粒子的辐射特性,分析了AVHRR数据确定的沙尘指数和地面气象站的能见度、风速等气象要素的关系,进而将沙尘指数进行了级别划分来确定沙尘暴的强度,得出沙尘暴强度的一种定量分析方法。结果表明,这种方法在利用气象卫星监测沙尘暴强度方面具有较好的效果。     

Development of a new dust index NDLI for Asian dust extraction system based on Aqua MODIS data and monitoring of trans-boundary Asian dust events in Japan

Heavy Asian dust events occur due to the strong wind in the Gobi deserts and are occasionally carried to Korea, Japan, and North America. They cause problems in human lives, such as respiratory diseases, transportation disturbances due to reduce visibility, and other disruptions in social activities. Remote sensing technology is useful for detecting and monitoring such airborne dust and understanding the distributions and movements of dust. To understand the Asian dust events, in this study, a new dust index is developed for the efficient detection of airborne Asian dust, which is a composite of two Moderate Resolution Imaging Spectroradiometer (MODIS) indices: Brightness Temperature Difference (BTD) and Normalized Difference Dust Index (NDDI). Our proposed Normalized Dust Layer Index (NDLI) detects dust more efficiently. To identify the characteristics of annual Asian dust events in Japan, a statistical time-series analysis of data from the years 2010, 2013 and 2014 is performed, and it is found that the dust events in 2014 were relatively calmer than those in 2013. An evaluation that was based on ground observations over different sites in Japan indicated that the proposed method performed well. Finally, we integrated our NDLI product into the trans-boundary air pollution satellite image database (TAPSIDB) system for monitoring Asian dust events.     

Remote sensing of mineral dust over land with MSG infrared channels: A new Bitemporal Mineral Dust Index

A new Bitemporal Mineral Dust Index (BMDI) is derived from Meteosat Second Generation (MSG) infrared observations over land at two different time slots per day.This daily dust index is evaluated with AErosol RObotic NETwork (AERONET) surface observations, MODerate resolution Imaging Spectro-radiometer (MODIS) “Deep Blue” Aerosol Optical Depth (AOD) and Ozone Monitoring Instrument (OMI) Aerosol Index, showing a good capability of the BMDI for dust detection and dust load estimation over land and also over deserts. BMDI dust detection is shown to be limited in scenes with high atmospheric humidity as e.g. coastal regions. In particular the insensitivity of BMDI to biomass burning aerosol is shown, leading to the possibility of remote sensing of mineral dust also in regions with large contributions of biomass burning aerosol to the total column aerosol concentrations. Time series of mineral dust as inferred from BMDI for the year 2006 are presented for four regions over the Sahara. These time series show strong (and different) annual cycles of dust load for all four regions. Especially the strong episodic character of atmospheric dust in the main source regions can be inferred from BMDI observations.     

深蓝算法应用于GF-1 16m相机反演陆地气溶胶

高分一号卫星是我国发展的新一代高分辨率对地观测卫星,如何应用该数据进行环境空气监测是目前迫切需要解决的问题。在深蓝算法基础上,根据GF-1星16 m相机的波段特征,借助MODIS的地表反射率产品去除地表贡献,从蓝波段数据反演了陆地气溶胶,实现了深蓝算法在GF-1星16m相机的应用。在此基础上,收集了2014年8~11月过境北京地区的GF-1星16m相机数据进行了反演实验,结果表明:该算法获取的气溶胶反演结果较好地反映了气溶胶的空间分布。同时,利用同期的AERONET/PHOTONS北京站的地面监测数据进行了算法验证,结果表明,本算法与地面数据有较好的相关性,相关系数大约为0.9,但该算法结果明显高于地面观测结果,可能是MODIS与GF-1星16m相机的波段响应不同导致的结果。     

地面能见度观测在卫星遥感IDDI指数中的融合应用

为了弥补卫星遥感沙尘监测云区之下的缺失信息,同时改进卫星遥感沙尘产品的精度,进行地面水平能见度资料与风云二号气象卫星红外差值沙尘指数(InfraRed Difference Dust Index,IDDI)的融合处理。首先采用2006~2010年3~5月沙尘多发期4:00~6:00的IDDI以及6:00的水平能见度建立两者的相关关系,结果表明:在西北地区,卫星沙尘指数IDDI与站点水平能见度之间呈较好的线性相关关系,而且3~5月的不同月份呈现不同的相关系数。然后利用所建立的相关关系将水平能见度资料转换为IDDI;最后对其进行融合处理,对存在IDDI观测值的像素利用周围站点的水平能见度资料采用反距离加权法进行修正,对不存在IDDI观测值的情况利用周围站点的水平能见度资料采用反距离加权法进行插值,得到沙尘观测结果。将该方法分别应用于2007年5月10日以及2008年4月20日发生在西北地区的沙尘过程中,得到较好的融合结果。同时显示出该方法更适用于沙尘过程强度较大,地面观测站点主要集中在卫星观测边缘附近的特征。