1. Introduction
1.1. Background
Greenhouse gas (GHG) emissions from agriculture and land-use contribute about one-quarter of global anthropogenic emissions (FAOSTAT 2016). Tropical subregions of Africa are major contributors of carbon dioxide (CO2), carbon monoxide (CO), black carbon (BC), and nitrogen dioxide (NO2) (Piketh et al. 1999; Jury 2000; Arino et al. 2012; Krotkov et al. 2016). Together with ozone (O3) and methane (CH4) emissions from biomass burning and livestock production, respectively, the GHG longwave radiation absorbers are likely to increase in the future (Dentener et al. 2005; Lamarque et al. 2010; Andreae and Ramanathan 2013). Hence, there is a need to better quantify trace gas concentrations over tropical Africa, as well as their sources and distribution in space and time. Emission estimates and inventories from rural agricultural sources have uncertainties that can be resolved via satellite observations and weather models incorporating dispersion. Unlike in situ point measurements, satellite observations provide useful information on regional-scale GHG concentrations (Bergamaschi et al. 2007; van der A et al. 2008; Marais et al. 2014a,b).
Ethiopia’s ~100 million people subsist on a gross domestic product of ~$54 billion (in 2015) and have low energy consumption (4290 MW) and small urban industrial emissions (World Bank 2016). Electricity does not reach 73% of the population, so rural households depend on gas and fuelwood for cooking and heating. Additionally, 88% of the population live in highlands above 1500 m, which covers 44% of the country (Desalegn et al. 2014; World Bank 2016), with mean minimum temperature <15°C (cf. Figure 1a). The United Nations Development Program (UNDP 2011) and the Food and Agriculture Organization Statistical Database (FAOSTAT 2016) GHG emission estimates are 50%–60% from agriculture, 19%–37% land-use, 3%–7% urban, 3%–6% waste, 3% transport, 2%–3% energy, and 2%–3% industry. Agricultural burning of ~30 billion kg covers ~4.6 Mha of Ethiopia (Giglio et al. 2003; cf. Figure 3a below). Ethiopia’s ~132 million livestock (in 2015) account for 20% of agriculture value and release significant quantities of GHG and aerosols (Teferi et al. 2013). Cattle comprise >50% of livestock (USAID 2013) and emit >30 kg of CH4 per annum per head (Broucek 2014). Forests cover <12% of the country, savannah >65%, and croplands >16% (FAS 2016). Soil bacteria and seasonal lightning release nitrates (Yienger and Levy 1995; Christian et al. 2003), but there are few volcanic or industrial sulfate sources. The dense vegetation, biomass burning, and livestock emit volatile organic compounds and CH4 that react to form fine aerosols, ozone, and formaldehyde (Marais et al. 2014a).
Mean maps averaged 2003–15 of (a) MODIS daytime land-surface temperature and index area (dashed), (b) MODIS color vegetation index, and (c) MERRA zonal circulation (W ×70) in vertical cross section averaged 6°–13°N with topography and calm zone U contoured. (d) Index area mean diurnal cycle of (left to right) MERRA hourly PBL height and surface wind speed and 3-hourly CMORPH rainfall. Local time is UTC + 3.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Mean maps averaged 2003–15 of (a) MODIS daytime land-surface temperature and index area (dashed), (b) MODIS color vegetation index, and (c) MERRA zonal circulation (W ×70) in vertical cross section averaged 6°–13°N with topography and calm zone U contoured. (d) Index area mean diurnal cycle of (left to right) MERRA hourly PBL height and surface wind speed and 3-hourly CMORPH rainfall. Local time is UTC + 3.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Mean maps averaged 2003–15 of (a) MODIS daytime land-surface temperature and index area (dashed), (b) MODIS color vegetation index, and (c) MERRA zonal circulation (W ×70) in vertical cross section averaged 6°–13°N with topography and calm zone U contoured. (d) Index area mean diurnal cycle of (left to right) MERRA hourly PBL height and surface wind speed and 3-hourly CMORPH rainfall. Local time is UTC + 3.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Satellites have provided estimates of atmospheric particulates since 1980 (Burrows et al. 2011). Recent advances in passive remote sensing, such as NASA’s Atmospheric Infrared Sounder (AIRS) and Ozone Monitoring Instrument (OMI), enable the mapping of near-surface GHG distributions (2003+). Trace gas estimation depends on ultraviolet absorption (λ: 0.31–0.46 μm) (Krotkov et al. 2006), while aerosols >0.1-μm diameter and carbon gases induce near-infrared absorption (λ: 0.5–7.6 μm). Satellite retrievals for Ethiopia must contend with seasonal obscuration by Saharan dust and thick clouds. Rural and biogenic emissions are diffused and may be insensitive to wind trajectory and planetary boundary layer (PBL) height, compared with their urban industrial counterparts. In addition to satellite retrievals, the Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2), reanalysis model (Molod et al. 2015) with air chemistry assimilation (Rienecker et al. 2016) helps quantify dispersion conditions over Ethiopia. Although little in situ data on urban industrial and agricultural emissions are available, source inventories provide adequate information (CAIT 2016).
1.2. Motivation and objectives
Harmful air constituents can limit productivity through human respiratory complications and reduced resource yields (van Wesenbeeck et al. 2009; Fantaye 2016). Ethiopia experiences a mortality rate of 57/100 000 from air pollution exposure (WHO 2016), with total pollution-related deaths of 129 000 in 2015 (Landrigan et al. 2018). The country is a major source of CH4 emissions (FAOSTAT 2016).
The central objective of this work is to characterize the near-surface air chemistry at daily to seasonal time scales by analysis of satellite- and model-estimated trace gases, particulates, and meteorological conditions over the Ethiopian highlands. Scientific questions include the following:
What are the characteristic patterns and seasonality of near-surface air constituents?
Which are regionally imported versus locally generated?
How are local emissions transported and dispersed?
Which meteorological factors affect Ethiopia’s air chemistry at annual and daily time scales?
2. Data and methods
Anticipating rural agricultural emissions from the Ethiopian highlands, our study of air chemistry covers 3°–17°N, 31.5°–48°E, and the subset “index” area for the time series is 5°–14°N, 35°–41.5°E (Figure 1a). The geographic focus excludes the lowlands surrounding Ethiopia, where vegetation fraction and agricultural emissions are half the highlands. Maps of MODIS fire counts (Justice et al. 2002), livestock density (Robinson et al. 2014), and lightning strikes (Cecil et al. 2015) yield insights on emission sources. Short-lived trace gas estimates from the OMI satellite include cloud-screened tropospheric NO2 (Bucsela et al. 2013; Lamsal et al. 2014) and (low level) PBL sulfur dioxide (SO2; Krotkov et al. 2006) interpolated to 25-km resolution from October 2004 to December 2015. Multilevel trace gas estimates of CH4, CO, CO2, and O3 derive from the AIRS satellite (Maddy et al. 2008; Xiong et al. 2008; Yurganov et al. 2008; Warner et al. 2010; Thonat et al. 2012; Levy et al. 2013) at 100-km resolution from September 2002 to December 2015, as daytime retrievals averaged in the 850–700-hPa layer. MISR satellite 0.55 μm aerosol optical depth (AOD) at 50-km resolution describes mineral dust intrusions from adjacent deserts. MERRA-2 air chemistry data (Molod et al. 2015) include monthly near-surface SO2, BC, dust particulates (<10 μm) and O3 concentration at ~50-km resolution from January 2000 to December 2015. The GEOS-5 system assimilates in situ and satellite observations, combined with emission inventories (Buchard et al. 2015; da Silva et al. 2015). The satellite inputs are incremental and start with tropospheric O3 and AOD in 1980 (Torres et al. 2002), CO in the 1990s, and tropospheric trace gases in the 2000s. The above references provide information on how the individual satellites retrieve air chemistry measurements. Algorithms that convert radiometer responses to measurements depend on in situ validations in a variety of environments. Similar work in the more industrialized South Africa (Jury 2017) has determined that satellite SO2 estimates are ambiguous, whereas other variables show agreement with in situ measurements and inventory-based model simulations.
The MERRA-2 meteorological data employed here include 850- and 775-hPa air temperature and lapse rate (dT/dz), 850–700-hPa wind components, 850–775-hPa relative divergence, 700-hPa specific humidity, 500-hPa geopotential height and vertical motion, and PBL height, calculated as the depth of surface-induced turbulence. “Near surface” averages in the 850–700-hPa layer represent transport and dispersion over the Ethiopian highlands. Additional data include satellite color vegetation fraction at 8-km resolution (NDVI; Tucker et al. 2005) and multisatellite estimated rainfall at 25-km resolution (CMORPH; Joyce et al. 2004). Table 1 summarizes the datasets, acronyms, and characteristics.
Datasets used in the analysis. All data are sourced from NASA Giovanni website, except as footnoted.
Time series of index area-averaged monthly and daily AIRS, MISR, and OMI satellite data and MERRA-2 model data were formulated. Gaps in daily satellite retrievals 2003–15 (from passive polar-orbiting radiometers) were in-filled by calculation of a 3-day running maximum. Statistical analyses utilized Pearson product–moment cross-correlation of air constituents and meteorology for the mean seasonal cycle and continuous daily fluctuations. Given the temporal filtering and serial autocorrelation, 192 months reduced to 4 degrees of freedom, and 4748 days reduced to ~30 degrees of freedom, yielding statistical thresholds for 90% confidence of |0.7| monthly and |0.3| daily. Comparisons of global and local averages were made for seasonal cycles of CO2 and CH4.
To understand the air chemistry concentration, transport, and dispersion, seasonal mean maps were calculated for BC, CH4, CO, 700-hPa winds, sinking motion, surface temperature, and rainfall. Vertical cross-sections of mean AIRS satellite CO, CH4, and O3 were analyzed up to 500 hPa, and the diurnal cycle of MERRA-2 hourly PBL height, near-surface wind speed, and 3-hourly CMORPH rainfall was calculated. Vertical cross-sections of the MERRA-2 meridional circulation and wind vorticity were analyzed for key features, such as the equatorial trough and subtropical jet.
The daily time series were ranked, and MERRA-2 near-surface winds were analyzed for the peak episode of 24–26 December 2010, when NO2 and other trace gas concentrations were high. The atmospheric dispersion of a continuous virtual source at 9°N, 37°E, 100 m in the period 24–26 December 2010 was simulated by the HYSPLIT model underpinned with GFS weather data and by radiosonde profiles from southwestern Arabia. Aerosol density cross sections were obtained on 24 December 2010 from the Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO; Winker et al. 2007). Each air constituent has certain units: MERRA-2 variables are kg m−3 (typically ×10−9); AIRS O3, CH4, and CO is ppm or μg m−3 (typically ×10−8); and OMI NO2 is cm−2 (×1016).
The research outcomes subdivide into (i) climatology and seasonal cycle, (ii) temporal statistics and relative influence, and (iii) daily air chemistry fluctuations and scenario of maxima. The scope of work excludes validations due to unavailability of local in situ measurements. Work elsewhere has sought to minimize uncertainties in satellite trace gas retrievals through algorithm refinement (Maddy et al. 2008; Xiong et al. 2008; Yurganov et al. 2008; Jury 2017). With regard to CH4, Bergamaschi et al. (2007) note ±1.5% uncertainty in satellite retrievals and ±3-ppb uncertainty in surface measurements. Frankenberg et al. (2006) found that “cloud screening” reduces systematic errors below 3% for satellite retrievals of constituents such as CH4, CO2, and NO2.
3. Results
3.1. Climatology and seasonal cycle
MODIS satellite data averaged 2003–15 (Figures 1a,b) characterize the highland climate: cool land temperatures (~25°C) and moderate vegetation fraction (~0.4). The mean zonal circulation (Figure 1c) reveals winds from the east with a rising component in the low levels and a sinking component in the upper levels. There is a wind shadow just west of the Rift Escarpment, where trace gases tend to accumulate. The mean diurnal cycle averaged over the highlands area (Figure 1d) displays large-amplitude changes in PBL height (120–1800 m) and rainfall (0.8–3.6 mm h−1) and modest changes of wind speed (4–5 m s−1). The daytime satellite overpass corresponds with a deep boundary layer prior to moist convection. Daytime near-surface winds tend to converge onto the highlands (cf. Figure 6c; also Jury 2016) and tend to limit (enhance) horizontal (vertical) dispersion.
The 2003–15 mean north–south and east–west cross-sections (Figures 2a–c) of AIRS CO, CH4, and O3 characterize the vertical structure of trace gases over Ethiopia, while emission sources (Figures 3a–c) provide context. CO (Figure 2a) displays a low-level maximum over the western highlands (~10°N, 36°E, 850 hPa) consistent with biomass burning (cf. Figure 3a) and westward transport, forming the African smoke plume (Giglio et al. 2003). CH4 (Figure 2b) also displays a low-level maximum over the highlands (~9°N, 39°E, 700 hPa) where the livestock are located (cf. Figure 3b). In contrast, O3 (Figure 2c) increases aloft and to the north, indicating photochemical production from more distant sources in the subtropics (Cooper et al. 2014). The mean climatology map for NO2 (Figure 3c) illustrates plumes extending westward from the highlands, attributable to biomass burning, lightning, and minor urban industrial sources (van der A et al. 2008). The NO2 plume emanating from the city of Addis Ababa (population 4.6 million in 2016) corresponds with in situ particulate concentrations up to 1 mg m−3 (Abate 2015; Tefera et al. 2016).
Mean 2003–15 vertical cross-sections (left: S–N avg 35°–41.5°E; right: W–E avg 6°–13°N) of AIRS daytime (a) CO, (b) CH4, and (c) O3, all in ppm with topography. The index area is dashed.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Mean 2003–15 vertical cross-sections (left: S–N avg 35°–41.5°E; right: W–E avg 6°–13°N) of AIRS daytime (a) CO, (b) CH4, and (c) O3, all in ppm with topography. The index area is dashed.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Mean 2003–15 vertical cross-sections (left: S–N avg 35°–41.5°E; right: W–E avg 6°–13°N) of AIRS daytime (a) CO, (b) CH4, and (c) O3, all in ppm with topography. The index area is dashed.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
(a) Map of MODIS fire counts summed over 2003–15 with 1500-m elevation contour. (b) Livestock density (number per square kilometer; adapted from Wint and Robinson 2007) with index area (dashed). (c) OMI satellite tropospheric NO2 climatology 2005–16 number per square centimeter) with LIS satellite lightning flash density >300 km−2 raster gray shaded. Largest city is shown as a circle.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
(a) Map of MODIS fire counts summed over 2003–15 with 1500-m elevation contour. (b) Livestock density (number per square kilometer; adapted from Wint and Robinson 2007) with index area (dashed). (c) OMI satellite tropospheric NO2 climatology 2005–16 number per square centimeter) with LIS satellite lightning flash density >300 km−2 raster gray shaded. Largest city is shown as a circle.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
(a) Map of MODIS fire counts summed over 2003–15 with 1500-m elevation contour. (b) Livestock density (number per square kilometer; adapted from Wint and Robinson 2007) with index area (dashed). (c) OMI satellite tropospheric NO2 climatology 2005–16 number per square centimeter) with LIS satellite lightning flash density >300 km−2 raster gray shaded. Largest city is shown as a circle.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
The Ethiopian highland mean seasonal cycle (Figures 4a–f) displays low amplitude for SO2, O3, PBL height, and V wind and high amplitude for other variables. The climatic insensitivity of SO2, O3, and dust (AOD) suggests inward transport from distant sources (Knippertz and Todd 2012). PBL height is flat because of the offsetting effect of moist convection during summer, while the north–south shift of the equatorial trough exerts minimal influence on V winds due to orography. Among those variables with large seasonal cycle, CH4 lags vegetation and maximizes in October–December, while the BC and CO peak in January–February with biomass burning. The Ethiopian area-averaged CH4 is above the global average CH4 from November to March (cf. Figure 4b). Meteorological variables with large but narrow seasonal range include rainfall and humidity (July–August peak). Wind convergence rises to a maximum in June, whereas westward transport (−U wind) displays a broad peak in the October–May dry season. Ethiopia CO2 values are 1–2 ppm below global due to small industrial emissions, whereas CH4 values are 2–3 ppm above global due to agricultural emissions.
Mean seasonal cycle (×2) within the index area of (a) MERRA air constituents, (b) AIRS methane and vegetation, (c) MERRA BC, AIRS CO, and SHF, (d) AIRS global and local CO2, (e) MERRA meteorological variables, and (f) CMORPH rainfall and near-surface wind divergence. Seasonal cycle correlations are listed in Table 2.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Mean seasonal cycle (×2) within the index area of (a) MERRA air constituents, (b) AIRS methane and vegetation, (c) MERRA BC, AIRS CO, and SHF, (d) AIRS global and local CO2, (e) MERRA meteorological variables, and (f) CMORPH rainfall and near-surface wind divergence. Seasonal cycle correlations are listed in Table 2.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Mean seasonal cycle (×2) within the index area of (a) MERRA air constituents, (b) AIRS methane and vegetation, (c) MERRA BC, AIRS CO, and SHF, (d) AIRS global and local CO2, (e) MERRA meteorological variables, and (f) CMORPH rainfall and near-surface wind divergence. Seasonal cycle correlations are listed in Table 2.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Among the seasonal cycle correlations in Table 2, AOD, dust, and CO2 are inversely related with CH4. The BC increases in dry weather, and CO follows sensible heat flux (SHF), indicating that drying is linked with biomass burning. The wind variables reveal transport and mixing influences: northerly winds (−V) increase CO2, southerlies increase BC concentrations, westerly winds bring dust (AOD), and easterlies (−U) raise CO, CH4, and O3 concentrations. These links suggest horizontal advection as the main driver. NDVI is negative correlated with SHF, CO, and PBL height, which rise in the dry season.
Correlation of the seasonal cycle of near-surface air constituents and meteorology averaged over the Ethiopia index area; values above 90% confidence with N = 12 are in bold font. Data identifications are as follows. r: MISR, m: MERRA-2, Dust: particulate matter <10 μm, a: AIRS, PBL: planetary boundary layer height, U and V: wind components, Humid: specific humidity, NDVI: vegetation, DIV: wind divergence, s: satellite, and SHF: sensible heat flux. Graphs are shown in Figure 4.
3.2. Seasonal features
The seasonal distribution of near-surface trace gases is illustrated in Figures 5a–c. Having outlined the mean seasonal cycle, both CO and BC show high concentration over the southwest highlands in the December–February season and low concentration during the wet season. The CH4 shows a persistent livestock “footprint” at higher elevations with a maximum at the end of the growing season (September–November) over the north-central highlands.
Seasonal mean maps of daytime near-surface (a) CO, (b) CH4, (c) O3, and (d) BC. Each column has the same scale, and maximum values are labeled; 1500-m elevation is contoured.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Seasonal mean maps of daytime near-surface (a) CO, (b) CH4, (c) O3, and (d) BC. Each column has the same scale, and maximum values are labeled; 1500-m elevation is contoured.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Seasonal mean maps of daytime near-surface (a) CO, (b) CH4, (c) O3, and (d) BC. Each column has the same scale, and maximum values are labeled; 1500-m elevation is contoured.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
The seasonal atmospheric circulation (Figures 6a–d) is dominated by upper-level anticyclonic vorticity (and subsidence) imparted by the subtropical jet stream, especially in the December–February season. By March–May, a shallow overturning circulation develops in the south as the subtropical jet weakens. Northeasterly flow shifts to easterly and surface air temperatures rise in March–May before the wet season. The equatorial trough and its cyclonic vorticity push northward to the highlands in the June–August season. Increasing rainfall and wet deposition reduces trace gas and aerosol concentrations, except for CH4. The meridional overturning circulation weakens in the September–November season, and near-surface easterly flow is reestablished.
(left) Seasonal mean vertical cross sections averaged 35°–41.5°E of wind vorticity (shaded ×10−5, anticyclonic negative orange), zonal wind (contours > 10 m s−1), and meridional circulation (vectors, W ×60) with topography; (lower left) index area bracket. (right) Seasonal mean maps of (850–700 hPa) near-surface winds, with same scale. Map overlays (a) DJF: 700-hPa sinking motion >0.045 Pa s−1 and flow channeling; (b) MAM: air temp >30°C; (c) JJA: CMORPH rainfall >5 mm day−1; and (d) SON: 1500-m elevation contour.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
(left) Seasonal mean vertical cross sections averaged 35°–41.5°E of wind vorticity (shaded ×10−5, anticyclonic negative orange), zonal wind (contours > 10 m s−1), and meridional circulation (vectors, W ×60) with topography; (lower left) index area bracket. (right) Seasonal mean maps of (850–700 hPa) near-surface winds, with same scale. Map overlays (a) DJF: 700-hPa sinking motion >0.045 Pa s−1 and flow channeling; (b) MAM: air temp >30°C; (c) JJA: CMORPH rainfall >5 mm day−1; and (d) SON: 1500-m elevation contour.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
(left) Seasonal mean vertical cross sections averaged 35°–41.5°E of wind vorticity (shaded ×10−5, anticyclonic negative orange), zonal wind (contours > 10 m s−1), and meridional circulation (vectors, W ×60) with topography; (lower left) index area bracket. (right) Seasonal mean maps of (850–700 hPa) near-surface winds, with same scale. Map overlays (a) DJF: 700-hPa sinking motion >0.045 Pa s−1 and flow channeling; (b) MAM: air temp >30°C; (c) JJA: CMORPH rainfall >5 mm day−1; and (d) SON: 1500-m elevation contour.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Considering the dry-season map (Figure 6a), a key feature is the horizontal diffluence of easterly flow reaching the Rift Escarpment. Subtropical troughs channel flow into the Red Sea and draw air from the northern highlands, while easterlies pass across the southern highlands and Turkana Valley. North–south bands of subsiding air create standing waves and a dry stable layer (dT/dz −5°C km−1). The Froude number Fr = U/N h (wind speed divided by buoyancy index and mountain height; Reisner and Smolarkiewicz 1994) falls below 0.3, so a wind shadow forms west of the Rift Escarpment and accumulates trace gases.
3.3. Intraseasonal variability
Daily time series of near-surface satellite trace gases are given in Figures 7a–c. Of concern is the rising trend of daily CH4 that accounts for a remarkable 43% of total variance related to the concomitant growth of human and livestock populations. The seasonal cycle is of similar amplitude to episodic fluctuations, suggesting steady emissions influenced more by slowly varying vegetation cover than dispersion conditions. In contrast, CO exhibits a weak decline and large episodic fluctuations (from biomass burning). NO2 episodes far exceed the seasonal cycle, particularly in the dry seasons of 2010–11 and 2014–15. Spikes in the NO2 record tend to grow with time and relate to weather conditions (cf. section 3.4.), yet the overall trend is minimal.
Daily time series of Ethiopia index area near-surface (a) AIRS CH4, (b) AIRS CO, and (c) OMI NO2, incorporating a 3-day running maximum. Linear trends are given in (a),(b); arrow points to case study in (c).
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Daily time series of Ethiopia index area near-surface (a) AIRS CH4, (b) AIRS CO, and (c) OMI NO2, incorporating a 3-day running maximum. Linear trends are given in (a),(b); arrow points to case study in (c).
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Daily time series of Ethiopia index area near-surface (a) AIRS CH4, (b) AIRS CO, and (c) OMI NO2, incorporating a 3-day running maximum. Linear trends are given in (a),(b); arrow points to case study in (c).
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Table 3 lists the daily correlations and shows interdependence of near-surface CO and NO2 (r = +0.50) and CO and satellite O3 (+0.70). The correlation of model versus satellite O3 is disappointing (+0.23), likely because retrievals are tuned for the upper troposphere. CO relates to humidity and dewpoint temperature (−0.52, −0.55 drier), lapse rate (+0.45 stable), and OLR (+0.44 warmer). NO2 mirrors the CO meteorology correlations; both show negative association with vorticity and U wind. Although anticyclonic easterly winds could draw air pollution from the Arabian Peninsula (Farahat 2016), most distribution maps (cf. Figures 2, 5) show low concentrations over the Red Sea and Gulf of Aden. Thus, we infer an accumulation of local emissions within the highlands wind shadow. CH4 shows relatively weaker meteorological influence, except for OLR (+0.37 warmer). Daily near-surface O3 from the model correlates with air temperature (+0.64 warmer), while O3 from the satellite relates to dry conditions. Figures 8a–f illustrate scatterplots of the leading air constituent–meteorology interrelationships are listed in Table 3. The CO scatterplot (cf. Figure 8b) is more coherent at low concentrations, corresponding with high humidity and wet deposition. A key feature is the insensitivity of methane concentrations to meteorological forcing.
Correlation of daily time series of near-surface air constituents and meteorology averaged over the Ethiopia index area; values above 90% confidence with N = 4748 are in bold font. Data identification areas follows. o: OMI, T: 775-hPa air temperature, dT/dz: lapse rate, OLR: satellite outgoing longwave radiation, Vort: wind vorticity (anticyclonic negative), and T dew: surface dewpoint temperature. Duplicate items are consistent with Table 2. Scatterplots are shown in Figure 8.
Scatterplots of daily time series comparing meteorology and air constituents averaged over the Ethiopia index area: (a) OLR vs CO, (b) specific humidity vs CO, (c) dT/dz vs CO, (d) air temperature vs O3, (e) OLR vs CH4, and (f) wind divergence vs CH4. Linear fit is given; daily correlations are listed in Table 3.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Scatterplots of daily time series comparing meteorology and air constituents averaged over the Ethiopia index area: (a) OLR vs CO, (b) specific humidity vs CO, (c) dT/dz vs CO, (d) air temperature vs O3, (e) OLR vs CH4, and (f) wind divergence vs CH4. Linear fit is given; daily correlations are listed in Table 3.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Scatterplots of daily time series comparing meteorology and air constituents averaged over the Ethiopia index area: (a) OLR vs CO, (b) specific humidity vs CO, (c) dT/dz vs CO, (d) air temperature vs O3, (e) OLR vs CH4, and (f) wind divergence vs CH4. Linear fit is given; daily correlations are listed in Table 3.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
3.4. Meteorology of trace gas episode
The NO2 episode of 24–26 December 2010 (Figures 9a–c) features dry anticyclonic easterly flow that slows abruptly on reaching the Rift Escarpment, creating a calm area over the central highlands of Ethiopia. The HYSPLIT simulation shows trace gas accumulation in the wind shadow to the west of the mountains. Radiosonde profiles over Arabia measured low dewpoint temperature (−30°C at 800 hPa) under westerly flow. CALIPSO aerosol sections and corresponding atmospheric circulation features during this event (Figure 9c) illustrate subsidence above the PBL along the edge of the subtropical jet stream. The top of the dense aerosol layer rises over the highlands but is constrained by meridional rotors on either side. The upper-level anticyclonic vorticity induces midlevel sinking motion and near-surface drying. The key feature of the episode is a deflection of stable easterly flow on reaching the Rift Escarpment. The consequent wind shadow acts like a magnet for trace gases from diverse sources.
(a) HYSPLIT simulated dispersion (shaded 0.1, 1, 10 × 10−6 kg m−3) 24–26 Dec 2010 from 9°N, 37°E, near-surface wind (vectors); dashed red line highlights the wind barrier. (b) Arabian radiosonde profiles of temperature, dewpoint, and wind at 0000 UTC 25 Dec 2010. (c) CALIPSO aerosol density at 1400 UTC 24 Dec 2010, with wind circulation (meridional vectors and zonal contours >10 m s−1) and topography overlay.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
(a) HYSPLIT simulated dispersion (shaded 0.1, 1, 10 × 10−6 kg m−3) 24–26 Dec 2010 from 9°N, 37°E, near-surface wind (vectors); dashed red line highlights the wind barrier. (b) Arabian radiosonde profiles of temperature, dewpoint, and wind at 0000 UTC 25 Dec 2010. (c) CALIPSO aerosol density at 1400 UTC 24 Dec 2010, with wind circulation (meridional vectors and zonal contours >10 m s−1) and topography overlay.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
(a) HYSPLIT simulated dispersion (shaded 0.1, 1, 10 × 10−6 kg m−3) 24–26 Dec 2010 from 9°N, 37°E, near-surface wind (vectors); dashed red line highlights the wind barrier. (b) Arabian radiosonde profiles of temperature, dewpoint, and wind at 0000 UTC 25 Dec 2010. (c) CALIPSO aerosol density at 1400 UTC 24 Dec 2010, with wind circulation (meridional vectors and zonal contours >10 m s−1) and topography overlay.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
4. Concluding discussion
This study advances our understanding of agricultural emissions that contribute to greenhouse warming using satellite estimates of key air constituents and, thus, has broad relevance to the wider scientific community.
The air chemistry of Ethiopia has been studied using seasonal and daily satellite estimates and inventory-based model products in the period 2000–16. CH4 reaches high concentrations due to livestock emissions over the north-central highlands and is above the global average from November to March (cf. Figure 4b). In contrast, CO rises over the western highlands in the dry season. Among the seasonal cycle correlations, dust, AOD, and CO2 are inversely related with CH4; BC and CO are negatively related with humidity and rainfall; and CO closely follows SHF, indicating that drier conditions favor biomass burning. Northerly (southerly) winds enhance CO2 (BC), and westerly winds enhance AOD, whereas easterly winds drive up CO, CH4, and O3. Hence, the slowly varying seasonal circulation and agricultural land-use underlie short-term weather fluctuations, which collectively determine trace gas concentrations.
Daily correlations reveal interdependence of CO and NO2 and of meteorological sensitivity to dry, stable weather conditions and anticyclonic easterly winds that create a calm area next to the Rift Escarpment. CH4 fluctuations are relatively insensitive to weather conditions, whereas near-surface O3 increases with air temperature. Dispersion conditions underlying high CO, CH4, and NO2 episodes are characterized by stable, dry, anticyclonic weather and midtropospheric subsidence from intrusion of the northern subtropical jet stream. The circulation and trace gas patterns suggest that many GHGs are locally generated, not imported. The wind shadow cast by the Rift Escarpment accumulates emissions such as CH4, which has increased >0.1 ppm from 2002 to 2016. Its linear upward trend accounts for 43% of variance and suggests the need to improve the efficiency of livestock production in Ethiopia. Future work will compare new in situ data with satellite retrievals.
Acknowledgments
The author gained valuable insights working in Ethiopia with Wollo University (appendix photo), Arba Minch University, and the Melkasa Institute for Agriculture Research, with support from the South African Department of Education. Some data analyses derive from NASA Giovanni, IRI Climate Library, APDRC Hawaii, NOAA ARL Ready, University of Wyoming radiosonde, and KNMI Climate Explorer. E. A. Marais of the University of Birmingham provided useful insights during preparation of the manuscript.
Appendix
Figure A1 is a photo of participants in the Wollo University air chemistry and meteorology data analysis workshop on 16 December 2016.
Participants in the Wollo University air chemistry and meteorology data analysis workshop, 16 Dec 2016.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Participants in the Wollo University air chemistry and meteorology data analysis workshop, 16 Dec 2016.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
Participants in the Wollo University air chemistry and meteorology data analysis workshop, 16 Dec 2016.
Citation: Earth Interactions 22, 12; 10.1175/EI-D-17-0009.1
References
Abate, Z., 2015: Impact of vehicle emission on air pollution in Addis Ababa. Stakeholder Workshop on Air Quality and Transportation Challenges in Ethiopia and Agenda for Clean Air Action Plan, Addis Ababa, Ethiopia, Ministry of Environment and Forest, The Federal Democratic Republic of Ethiopia and Centre for Science and Environment, 8 pp., www.cseindia.org/userfiles/ethiopia-report.pdf.
Andreae, M. O., and V. Ramanathan, 2013: Climate’s dark forcings. Science, 340, 280–281, https://doi.org/10.1126/science.1235731.
Arino, O., S. Casadio, and D. Serpe, 2012: Global night-time fire season timing and fire count trends using the ATSR instrument series. Remote Sens. Environ., 116, 226–238, https://doi.org/10.1016/j.rse.2011.05.025.
Bergamaschi, P., and Coauthors, 2007: Satellite chartography of atmospheric methane from SCIAMACHY on board ENVISAT: 2. Evaluation based on inverse model simulations. J. Geophys. Res., 112, D02304, https://doi.org/10.1029/2006JD007268.
Broucek, J., 2014: Production of methane emissions from ruminant husbandry: A review. J. Environ. Prot., 5, 51796, https://doi.org/10.4236/jep.2014.515141.
Buchard, V., and Coauthors, 2015: Using the OMI aerosol index and absorption aerosol optical depth to evaluate the NASA MERRA Aerosol Reanalysis. Atmos. Chem. Phys., 15, 5743–5760, https://doi.org/10.5194/acp-15-5743-2015.
Bucsela, E. J., and Coauthors, 2013: A new stratospheric and tropospheric NO2 retrieval algorithm for nadir-viewing satellite instruments: Applications to OMI. Atmos. Meas. Tech., 6, 2607–2626, https://doi.org/10.5194/amt-6-2607-2013.
Burrows, J. P., U. Platt, and P. Borrell, Eds., 2011: The Remote Sensing of Tropospheric Composition from Space. Springer-Verlag, 536 pp.
CAIT, 2016: Country GHG emissions. World Resources Institute Agricultural Emission Database, CAIT Climate Data Explorer, https://cait.wri.org/profile/Ethiopia.
Cecil, D. J., D. E. Buechler, and R. J. Blakeslee, 2015: TRMM LIS climatology of thunderstorm occurrence and conditional lightning flash rates. J. Climate, 28, 6536–6547, https://doi.org/10.1175/JCLI-D-15-0124.1.
Christian, H. J., and Coauthors, 2003: Global frequency and distribution of lightning as observed from space by the Optical Transient Detector. J. Geophys. Res., 108, 4005, https://doi.org/10.1029/2002JD002347.
Cooper, O. R., and Coauthors, 2014: Global distribution and trends of tropospheric ozone: An observation-based review. Elem. Sci. Anthropocene, 2, 29, https://doi.org/10.12952/journal.elementa.000029.
da Silva, A. M., C. A. Randles, V. Buchard, A. Darmenov, P. R. Colarco, and R. Govindaraju, 2015: The MERRA Aerosol Reanalysis. GMAO Tech. Rep. 7, 30 pp.
Dentener, F., D. Stevenson, J. Cofala, R. Mechler, M. Amann, P. Bergamaschi, F. Raes, and R. Derwent, 2005: The impact of air pollutant and methane emission controls on tropospheric ozone and radiative forcing: CTM calculations for the period 1990–2030. Atmos. Chem. Phys., 5, 1731–1755, https://doi.org/10.5194/acp-5-1731-2005.
Desalegn, T., F. Cruz, M. Kindu, M. B. Turrión, and J. Gonzalo, 2014: Land-use/land-cover (LULC) change and socioeconomic conditions of local communities in the central highlands of Ethiopia. Int. J. Sustain. Dev. World Ecol., 21, 406–413, https://doi.org/10.1080/13504509.2014.961181.
Fantaye, K. T., 2016: Description of cropping systems, climate, and soils in Ethiopia. Global Yield Gap Atlas, www.yieldgap.org/ethiopia.
FAOSTAT, 2016: Ethiopia. Food and Agriculture Organization of the United Nations, accessed December 2016, http://www.fao.org/faostat/en/#country/238.
Farahat, A., 2016: Air pollution in the Arabian Peninsula (Saudi Arabia, the United Arab Emirates, Kuwait, Qatar, Bahrain, and Oman): Causes, effects, and aerosol categorization. Arab. J. Geosci., 9, 196, https://doi.org/10.1007/s12517-015-2203-y.
FAS, 2016: Eastern Africa. U.S. Foreign Agriculture Service, www.pecad.fas.usda.gov/cropexplorer/imageview.aspx?regionid=eafrica.
Frankenberg, C., and Coauthors, 2006: Satellite chartography of atmospheric methane from SCIAMACHY on board ENVISAT: Analysis of the years 2003 and 2004. J. Geophys. Res., 111, D07303, https://doi.org/10.1029/2005JD006235.
Giglio, L., J. Descloitres, C. O. Justice, and Y. J. Kaufman, 2003: An enhanced contextual fire detection algorithm for MODIS. Remote Sens. Environ., 87, 273–282, https://doi.org/10.1016/S0034-4257(03)00184-6.
Joyce, R. J., J. E. Janowiak, P. A. Arkin, and P. P. Xie, 2004: CMORPH: A method that produces global precipitation estimates from passive microwave and infrared data at high spatial and temporal resolution. J. Hydrometeor., 5, 487–503, https://doi.org/10.1175/1525-7541(2004)005<0487:CAMTPG>2.0.CO;2.
Jury, M. R., 2000: The dry season climate of tropical southern Africa and implications for pyrogenic emissions. S. Afr. J. Sci., 96, 387–390.
Jury, M. R., 2016: Large-scale features of Africa’s diurnal climate. Phys. Geogr., 37, 120–131, https://doi.org/10.1080/02723646.2016.1163004.
Jury, M. R., 2017: Statistics and meteorology of air pollution episodes over the South African Highveld based on satellite–model datasets. J. Appl. Meteor. Climatol., 56, 1583–1594, https://doi.org/10.1175/JAMC-D-16-0354.1.
Justice, C. O., and Coauthors, 2002: The MODIS fire products. Remote Sens. Environ., 83, 244–262, https://doi.org/10.1016/S0034-4257(02)00076-7.
Knippertz, P., and M. C. Todd, 2012: Mineral dust aerosols over the Sahara: Meteorological controls on emission and transport and implications for modeling. Rev. Geophys., 50, RG1007, https://doi.org/10.1029/2011RG000362.
Krotkov, N. A., S. Carn, A. Krueger, P. Bhartia, and K. Yang, 2006: Band residual difference algorithm for retrieval of SO2 from the Aura Ozone Monitoring Instrument (OMI). IEEE Trans. Geosci. Remote Sens., 44, 1259–1266, https://doi.org/10.1109/TGRS.2005.861932.
Krotkov, N. A., and Coauthors, 2016: Aura OMI observations of regional SO2 and NO2 pollution changes from 2005 to 2015. Atmos. Chem. Phys., 16, 4605–4629, https://doi.org/10.5194/acp-16-4605-2016.
Lamarque, J.-F., and Coauthors, 2010: Historical (1850–2000) gridded anthropogenic and biomass burning emissions of reactive gases and aerosols: Methodology and application. Atmos. Chem. Phys., 10, 7017–7039, https://doi.org/10.5194/acp-10-7017-2010.
Lamsal, L. N., and Coauthors, 2014: Evaluation of OMI operational standard NO2 column retrievals using in situ and surface-based NO2 observations. Atmos. Chem. Phys., 14, 11 587–11 609, https://doi.org/10.5194/acp-14-11587-2014.
Landrigan, P. J., and Coauthors, 2018: The Lancet Commission on pollution and health. Lancet, 391, 462–512, https://doi.org/10.1016/S0140-6736(17)32345-0.
Levy, R. C., S. Mattoo, L. A. Munchak, L. A. Remer, A. M. Sayer, F. Patadia, and N. C. Hsu, 2013: The Collection 6 MODIS aerosol products over land and ocean. Atmos. Meas. Tech., 6, 2989–3034, https://doi.org/10.5194/amt-6-2989-2013.
Maddy, E. S., C. D. Barnet, M. Goldberg, C. Sweeney, and X. Liu, 2008: CO2 retrievals from the Atmospheric Infrared Sounder: Methodology and validation. J. Geophys. Res., 113, D11301, https://doi.org/10.1029/2007JD009402.
Marais, E. A., D. J. Jacob, A. Guenther, K. Chance, T. P. Kurosu, J. G. Murphy, C. E. Reeves, and H. O. Pye, 2014a: Improved model of isoprene emissions in Africa using Ozone Monitoring Instrument (OMI) satellite observations of formaldehyde: Implications for oxidants and particulate matter. Atmos. Chem. Phys., 14, 7693–7703, https://doi.org/10.5194/acp-14-7693-2014.
Marais, E. A., and Coauthors, 2014b: Anthropogenic emissions in Nigeria and implications for atmospheric ozone pollution: A view from space. Atmos. Environ., 99, 32–40, https://doi.org/10.1016/j.atmosenv.2014.09.055.
Molod, A., L. Takacs, M. Suarez, and J. Bacmeister, 2015: Development of the GEOS-5 atmospheric general circulation model: Evolution from MERRA to MERRA2. Geosci. Model Dev., 8, 1339–1356, https://doi.org/10.5194/gmd-8-1339-2015.
Piketh, S. J., H. J. Annegarn, and P. D. Tyson, 1999: Lower tropospheric aerosol loadings over South Africa: The relative contribution of aeolian dust, industrial emissions, and biomass burning. J. Geophys. Res., 104, 1597–1607, https://doi.org/10.1029/1998JD100014.
Reisner, J. M., and P. K. Smolarkiewicz, 1994: Thermally forced low Froude number flow past three-dimensional obstacles. J. Atmos. Sci., 51, 117–133, https://doi.org/10.1175/1520-0469(1994)051<0117:TFLFNF>2.0.CO;2.
Rienecker, M. M., and Coauthors, 2016: The GEOS-5 data assimilation system—Documentation of versions 5.0.1, 5.1.0, and 5.2.0. NASA Tech. Rep. NASA/TM-2007-104606, https://gmao.gsfc.nasa.gov/pubs/docs/GEOS-5.0.1_Documentation_r3.pdf.
Robinson, T. P., and Coauthors, 2014: Mapping the global distribution of livestock. PLOS ONE, 9, e96084, https://doi.org/10.1371/journal.pone.0096084.
Tefera, W., A. Asfaw, F. Gilliland, A. Worku, M. Wondimagegn, A. Kumie, J. Samet, and K. Berhane, 2016: Indoor and outdoor air pollution-related health problems in Ethiopia: Review of related literature. Ethiop. J. Health Dev., 30, 1–12.
Teferi, E., W. Bewket, S. Uhlenbrook, and J. Wenninger, 2013: Understanding recent land use and land cover dynamics in the source region of the Upper Blue Nile, Ethiopia: Spatially explicit statistical modeling of systematic transitions. Agric. Ecosyst. Environ., 165, 98–117, https://doi.org/10.1016/j.agee.2012.11.007.
Thonat, T., C. Crevoisier, N. A. Scott, A. Chédin, T. Schuck, R. Armante, and L. Crépeau, 2012: Retrieval of tropospheric CO column from hyperspectral infrared sounders—Application to four years of Aqua/AIRS and MetOp-A/IASI. Atmos. Meas. Tech., 5, 2413–2429, https://doi.org/10.5194/amt-5-2413-2012.
Torres, O., P. K. Bhartia, J. R. Herman, A. Sinyuk, P. Ginoux, and B. Holben, 2002: A long-term record of aerosol optical depth from TOMS observations and comparison to AERONET measurements. J. Atmos. Sci., 59, 398–413, https://doi.org/10.1175/1520-0469(2002)059<0398:ALTROA>2.0.CO;2.
Tucker, C. J., J. E. Pinzon, M. E. Brown, D. A. Slayback, E. W. Pak, R. Mahoney, E. F. Vermote, and N. El Saleous, 2005: An extended AVHRR 8-km NDVI dataset compatible with MODIS and SPOT vegetation NDVI data. Int. J. Remote Sens., 26, 4485–4498, https://doi.org/10.1080/01431160500168686.
UNDP, 2011: Ethiopia’s climate-resilient green economy: Green economy strategy. UNDP Tech. Rep., 200 pp., www.undp.org/content/dam/ethiopia/docs/Ethiopia%20CRGE.pdf.
USAID, 2013: Agricultural growth project—Livestock market development. USAID Rep. AID-663-C-12-00009, 160 pp., https://www.usaid.gov/sites/default/files/documents/1860/AGP-LMD%20Value%20Chain%20Analysis.pdf.
van der A, R. J., H. J. Eskes, K. F. Boersma, T. P. C. van Noije, M. Van Roozendael, I. De Smedt, D. H. M. U. Peters, and E. W. Meijer, 2008: Trends, seasonal variability and dominant NOx source derived from a ten year record of NO2 measured from space. J. Geophys. Res., 113, D04302, https://doi.org/10.1029/2007JD009021.
van Wesenbeeck, C. F. A., M. A. Keyzer, and M. Nubé, 2009: Estimation of undernutrition and mean calorie intake in Africa: Methodology, findings and implications. Int. J. Health Geogr., 8, 37, https://doi.org/10.1186/1476-072X-8-37.
Warner, J. X., Z. Wei, L. L. Strow, C. D. Barnet, L. C. Sparling, G. Diskin, and G. Sachse, 2010: Improved agreement of AIRS tropospheric carbon monoxide products with other EOS sensors using optimal estimation retrievals. Atmos. Chem. Phys., 10, 9521–9533, https://doi.org/10.5194/acp-10-9521-2010.
WHO, 2016: WHO global urban ambient air pollution database. WHO, accessed December 2016, www.who.int/phe/health_topics/outdoorair/databases/cities/en/.
Winker, D. M., W. M. Hunt, and M. J. McGill, 2007: Initial performance assessment of CALIOP. Geophys. Res. Lett., 34, L19803, https://doi.org/10.1029/2007GL030135.
Wint, W., and T. Robinson, Eds., 2007: Gridded Livestock of the World (GLW). Food and Agriculture Organization Rep., 131 pp.
World Bank, 2016: Ethiopia statistical database. The World Bank, accessed December 2016, https://data.worldbank.org/country/Ethiopia.
Xiong, X., C. Barnet, E. Maddy, C. Sweeney, X. Liu, L. Zhou, and M. Goldberg, 2008: Characterization and validation of methane products from the Atmospheric Infrared Sounder (AIRS). J. Geophys. Res., 113, G00A01, https://doi.org/10.1029/2007JG000500.
Yienger, J. J., and H. Levy, 1995: Empirical model of global soil-biogenic NOx emissions. J. Geophys. Res., 100, 11 447–11 464, https://doi.org/10.1029/95JD00370.
Yurganov, L. N., W. W. McMillan, A. V. Dzhola, E. I. Grechko, N. B. Jones, and G. R. van der Werf, 2008: Global AIRS and MOPITT CO measurements: Validation, comparison, and links to biomass burning variations and carbon cycle. J. Geophys. Res., 113, D09301, https://doi.org/10.1029/2007JD009229.
