Tropical and South Atlantic marine ecosystems support fisheries that have vital environmental and socioeconomic importance. In this Review, we outline how the El Niño–Southern Oscillation — a Pacific mode of sea surface temperature variability — influences Atlantic fisheries via teleconnections and cascading linkages between physical, biogeochemical and ecological systems. Connections are driven by tropical pathways (involving changes in atmospheric stability associated with the Walker circulation and tropospheric warming) and extratropical pathways (involving the Pacific–South American and Pacific–North American teleconnection patterns). Depending on the location, these pathways modify rainfall and river discharge, winds and upwelling, or a combination of both, impacting salinity, nutrient availability, primary production and, thus, fish recruitment, biomass and catch. Fishery responses are strongly species dependent, reflecting variations in behaviour between species to environmental factors (such as temperature, oxygen, salinity, habitat and food availability). This regional variability and species dependency, coupled with strong non-stationarity, highlights the complexity of El Niño–Southern Oscillation impacts on Atlantic marine ecosystems. This historical signal is projected to weaken in the future. Enhanced observational systems and refined ecosystem models are urgently needed to enhance predictive capabilities, reduce societal impacts and improve sustainable management in these regions. The influence of the El Niño–Southern Oscillation on Atlantic marine systems and fisheries is complex. This Review outlines the mechanisms by which El Niño–Southern Oscillation impacts the tropical and South Atlantic, connecting physical climate perturbations to biogeochemical and ecological responses.

The coastal regions off Angola and Namibia are renowned for their highly productive marine ecosystems in the southeast Atlantic. In recent decades, these regions have undergone significant long-term changes. In this study, we investigate the variability of these long-term changes throughout the annual cycle and explore the underlying mechanisms using a 34-year (1982–2015) regional ocean model simulation. The results reveal a clear seasonal dependence of sea surface temperature (SST) trends along the Angolan and Namibian coasts, with alternating positive and negative trends. The long-term warming trend in the Angolan coastal region is mainly explained by a pronounced warming trend in the austral spring and summer (November-January), while the decadal trend off Namibia results from a counterbalance of an austral winter cooling trend and an austral summer warming trend. A heat budget analysis of the mixed-layer temperature variations shows that these changes are explained by a long-term modulation of the coastal currents. The Angolan warming trend is mainly explained by an intensification of the poleward coastal current, which transports more warm equatorial waters towards the Angolan coast. Off Namibia, the warming trend is attributed to a reduction in the northwestward Benguela Current, which advects cooler water from the south to the Namibian coast. These changes in the coastal current are associated with a modulation of the seasonal coastal trapped waves that are remotely-forced along the equatorial waveguide. These long-term changes may have significant implications for local ecosystems and fisheries.

The intertropical convergence zone (ITCZ), with its twice-annual passage over central Africa, is considered as the main driver of the rainfall seasonality. In this ITCZ paradigm, high rainfall occurs over regions of large low-level convergence. But recently, this paradigm was challenged over central Africa. Here, we show that a shallow meridional overturning circulation-driven by surface conditions-plays a thermodynamical control on the rainfall seasonality over central Africa. Indeed, due to the local evaporative cooling effect, the foot of the ascending branch of Hadley cells occurs where the temperature is the warmest, indicating a thermal low. This distorts the southern Hadley cell by developing its bottom-heavy structure. As result, both shallow and deep Hadley cells coexist over central Africa year-round. The deep mode is associated with the poleward transport of atmospheric energy at upper levels. The shallow mode is characterized by a shallow meridional circulation, with its moisture transport vanishing and converging in the midtroposphere rather than at lower troposphere. This midtropospheric moisture convergence is also the dominant component that shapes the vertically integrated moisture flux convergence, with little contribution of African easterly jets. This convergence zone thus controls the precipitating convection. Its meridional migration highlights the interhemispheric rainfall contrast over central Africa and outlines the unimodal seasonality. On the other hand, forced by the Congo basin cell, the precipitable water regulates the deep convection from the vegetated surface of Congo basin, acting as a continental sea. This nonlinear mechanism separates the rainfall into three distinct regimes: the moisture-convergence-controlled regime, with convective rainfall exclusively occurring in the rainy season; the local evaporation-controlled regime with drizzle in the dry season; and the precipitable-water-controlled regime, with exponential rainfall increase in the dry season.

AbstractSouth of 15°S, southern Africa has a subtropical climate, which is affected by temperate and tropical weather systems and comes under the influence of the Southern Hemisphere high-pressure systems. Most rainfall occurs in austral summer, but the southwest experiences winter rainfall. Much of the precipitation in summer is of convective origin forced by large-scale dynamics. There is a marked diurnal cycle in rainfall in summer. The El Niño Southern Oscillation (ENSO) influences interannual rainfall variability. In austral summer, drought tends to occur during El Niño, while above-normal rainfall conditions tend to follow La Niña. During El Niño, higher than normal atmospheric pressure anomalies, detrimental to rainfall, occur due to changes in the global atmospheric circulation. This also weakens the moisture transport from the Indian Ocean to the continent. The opposite mechanisms happen during La Niña. On top of the variability related to ENSO, the Pacific Ocean also influences the decadal variability of rainfall. Additionally, the Angola Current, the Agulhas Current, the Mozambique Channel and the southwest Indian Ocean affect rainfall variability. Over the last 40 to 60 years, near-surface temperatures have increased over almost the whole region, summer precipitation has increased south of 10°S, and winter precipitation has mostly decreased in South Africa. Meanwhile, the Agulhas Current and the Angola Current have warmed, and the Benguela Current has cooled.

Large body of works has studied the impacts of different drivers on southern Africa precipitation. But the influence of the Kalahari thermal low (KTL) on the southern Africa precipitation variability is overlooked to date and deserve more attention. Here, we reconcile two most important features of the southern Africa climate – the Angola low and the Botswana high through the Kalahari thermal low. Results show that the relationship between the KTL and the Angola low has two distinct behaviors at different timescales: (i) at seasonal timescale (annual cycle), the Angola low, driven by the KTL, is also a thermal low; (ii) but at interannual timescale, the reversal of the relationship suggests that the Angola low behaves as a tropical low in summer and forms as a Gill-like response to the diabatic heating related to KTL. Secondly, the KTL shows a strong negative correlation with southern Africa rainfall, implying that a strong KTL reduces the rainfall intensity. This teleconnection can be explained by the leading mode of variability of KTL. However, the composite analysis highlights two physical mechanisms responsible of this teleconnection: the heat dome mechanism during the warm phase of KTL, conducive to below normal rainfall conditions over southern Africa. In the cold phase of KTL, the Angola low plays a crucial role in modulating the regional precipitation and leads to above normal rainfall conditions. Furthermore, our findings found that current climate models from CMIP6 are unable to reproduce the observed mechanisms.

Coastal zones with dense populations, low elevations and/or inadequate adaptive capacity are on the frontline of unprecedented impacts from climate change. The Gulf of Guinea (GoG), stretching from Liberia to Gabon, is in particular vulnerable to coastal flooding caused by local and/or climate-induced sea level rise. In this region, interannual to decadal coastal sea level changes remain poorly understood, mainly due to a lack of tide gauge stations. Here we use nearly three decades (1993–2021) of satellite altimetry data to study the link between the Equatorial Atlantic and coastal GoG sea level variability. The rate of mean sea level rise increased from 3.47 to 3.89 ± 0.10 mm/yr from the Equatorial oceanic domain to the GoG coastal area, with an acceleration of 0.094 ± 0.050 mm/yr2. This corresponds to a mean sea level rise of about 8.9 cm over the entire altimetry period, 1993–2021. We focus on the (extreme) warm/cold events that occur in both the GoG during Atlantic Niños, and along the Angola-Namibia coast during Benguela Niños. Both events are driven by remote forcing via equatorial Kelvin waves and local forcing by local winds, freshwater fluxes and currents intensifications. Analysis of altimetry-based sea level, sea surface temperature anomalies, 20 °C isotherm based PIRATA moorings, and the Argo-based steric and thermometric sea level allows us to follow the coastal trapped waves (CTWs) along the GoG, and its link with major events observed along the strong Equatorial Atlantic warmings in 2010, 2012, 2019 and 2021. Both 2019 and 2021 warming have been identified as the warmest event ever reported in this region during the last 40 years. A lag of 1 month is observed between equatorial and West African coastal trapped wave propagation. This observation may help to better anticipate and manage the effects of extreme events on local ecosystems, fisheries, and socio-economic activities along the affected coastlines. In order to enable informed decision-making and guarantee the resilience of coastal communities in the face of climate change, it emphasises the significance of ongoing study in this field.

Extreme events contribute significantly to rainfall variability in semi-arid regions like South Africa. Here, following the definition of a novel typology of rainfall extremes, disentangling large- and small-scale events in Part I, we use quality-controlled observational databases in South Africa, the ERA5 reanalysis and satellite estimates TRMM-3B42 to examine the relationship between these two types of rainfall extremes and different modes of climate variability at various timescales. At low frequencies, rainfall extremes are assessed at interannual (IV: 2-8 years) and quasi-decadal (QDV: 8-13 years) timescales, which are primarily associated with the El Nino-Southern Oscillation (ENSO) and the Interdecadal Pacific Oscillation (IPO), respectively. At subseasonal timescales, the typology of rainfall extremes is analysed depending on the synoptic configurations, as inferred by seven convective regimes including tropical temperate troughs (TTTs: 3-7 days), and the intraseasonal variability associated with the Madden-Julien Oscillation (MJO: 30-60 days). At the IV timescale, the occurrence of large-scale extremes is substantially higher during its wet phases thereby suggesting a 400% rise in the occurrence of large-scale extremes as compared to its dry phases. At the QDV timescale, variability mostly relates to the modulation of small-scale extremes during its wet phases. Teleconnections with global sea surface temperature (SST) confirm that La Nina conditions favour overall wet conditions and extremes in South Africa. The numbers of large-scale extremes are consistently related to warmer SSTs in the North Atlantic, while their link with warmer Indian and tropical South Atlantic oceans is found to be statistically independent of the state of ENSO. At the subseasonal timescales, large-scale extremes largely occur during three out of the seven convective regimes identified in the southern African region whereas small-scale extremes are nearly equiprobable during all convective regimes. The occurrence of large-scale extremes during continent-rooted TTT is further enhanced during the locally wet phases of the MJO and is symmetrically weakened during its dry phases.

In this paper, we review observational and modelling results on the upwelling in the tropical Atlantic between 10 degrees N and 20 degrees S. We focus on the physical processes that drive the seasonal variability of surface cooling and the upward nutrient flux required to explain the seasonality of biological productivity. We separately consider the equatorial upwelling system, the coastal upwelling system of the Gulf of Guinea and the tropical Angolan upwelling system. All three tropical Atlantic upwelling systems have in common a strong seasonal cycle, with peak biological productivity during boreal summer. However, the physical processes driving the upwelling vary between the three systems. For the equatorial regime, we discuss the wind forcing of upwelling velocity and turbulent mixing, as well as the underlying dynamics responsible for thermocline movements and current structure. The coastal upwelling system in the Gulf of Guinea is located along its northern boundary and is driven by both local and remote forcing. Particular emphasis is placed on the Guinea Current, its separation from the coast and the shape of the coastline. For the tropical Angolan upwelling, we show that this system is not driven by local winds but instead results from the combined effect of coastally trapped waves, surface heat and freshwater fluxes, and turbulent mixing. Finally, we review recent changes in the upwelling systems associated with climate variability and global warming and ad-dress possible responses of upwelling systems in future scenarios.

<p>The Intertropical Convergence Zone (ITCZ), with its twice-annual passage over central Africa, is considered as the main driver of the rainfall seasonality. But recently, this paradigm was challenged. To find out what are the main drivers of the annual cycle of rainfall over central Africa, we present a simple comprehensive paradigm with both local forcings and regional-scale processes playing crucial role. Due to the local evaporative cooling effect, the foot of the ascending branch of Hadley cells occurs where the temperature is the warmest, indicating a thermal low. This distorts the southern Hadley cell by developing its bottom-heavy structure. As result, both shallow and deep Hadley cells coexist over central Africa year–round. The deep mode is associated with poleward branches at upper levels that transport the atmospheric energy. The shallow mode is characterized by a meridional return flow in the mid-troposphere that transports the water vapour instead of lower branches as widely reported. This favours the building-up of the mid-tropospheric moisture flux convergence with a limited contribution of the midlevel easterly jet, conducive to deep convection. Embedded in this strong rising branch of Hadley cells at midlevels, the intense convective rainfall, and with it the rainfall maximum position, is seasonally controlled by the dynamics of the midlevel shallow meridional return flow. This highlights the interhemispheric rainfall contrast over central Africa and outlines its unimodal seasonality. On the other hand, forced by the Congo basin cell, the precipitable water regulates the deep convection from the vegetated surface of Congo basin, acting as a continental sea. This nonlinear mechanism separates the rainfall into three distinct regimes – (i) the moisture-convergence-controlled regime, with convective rainfall exclusively occurring in the rainy season and (ii) the local evaporation-controlled regime with drizzle and (iii) the precipitable-water-controlled regime, with exponential increase of rainfall that both occur during the dry season.</p>

Equatorward decrease of mesoscale eddy abundance, and meridional variations of their mean properties in the tropical South-East Atlantic Ocean have been investigated on the basis of horizontal length-scales of geostrophic flows, and theories of two-dimensional geostrophic turbulence on a fl-plane. Meridional variations of satellite-derived altimetric eddy kinetic energy density spectra were assessed in bins of 5 degrees, from 5 degrees S to 25 degrees S, zonally averaged between 8 degrees W and 15 degrees E. They revealed two latitudinal bands of minimum eddy energy spectra: one in the northern Angola Basin (5 degrees S-10 degrees S) and the other across the Angola Benguela Front (15 degrees S-20 degrees S). In the latter, both the eddy forcing scale and the Rhines scale were observed at wavelengths of 300 and 350 km respectively, suggesting that only a small spectral range exists for eddies to grow. On the other hand, in the former, the eddy forcing scale and the Rhines scale were observed at wavelengths of 454 and 520 km respectively. The ratio between the Rhines scale and the first baroclinic Rossby radius revealed an onset of a critical latitude at about 11.4 degrees S. Consistently, maps of eddy frequency of occurrence computed from eddies identified in the region has shown a notable decay in signal of the eddy density distribution around that latitude. This could explain the sharp equatorward transition of mesoscale processes observed in this region, which seems to shift from nonlinear eddies to linear Rossby waves, potentially explaining why fewer eddies are observed in the region when compared to other regions of the world oceans.

The Angola Benguela Front (ABF), is a very dynamic area, characterized by a high-temperature gradient of up to 4°C per degree latitude. It fluctuates in position and intensity seasonally which strongly affects the local marine ecosystem. A lot of research, in the past decades, has focused on the SST variability at the interannual timescale in the ABF and the Angolan and Northern Namibian coast to the north and south of it in the contest of Benguela Niños and Niñas. A warming trend since the 1980’s in that region has been reported in the literature and was attributed to a decreasing trend in wind speed. In this study, we look at the processes responsible for the warming in the ABF region. The OGCM NEMO model is used for that matter. The results suggest that the warming is due to various processes acting during different seasons. In autumn, the modelled SST warming trend occurs along the Angolan sector and it is associated with a positive trend in net surface heat flux (Qnet) and with the weakening of the vertical flow associated with the upwelling of cooler water to the surface. In early summer (November-January), the modelled SST warming trend occurs along the Angolan and Namibian sector and it is primarily associated with the intensification of a coastal poleward flow bringing more warm water from the tropics into the ABF region and with the weakening of vertical flow, while locally, Qnet trend generates a cooling trend. The modelled SST cooling trend that occurred south of the ABF, especially in winter and early spring, is primarily associated with a northwards trend in the horizontal subsurface current that advects cooler water from the south and an intensification of the upwelling of cold water to the surface.

The impact of El Nino-Southern Oscillation (ENSO) on the southern African climate is well documented and provides skill in the seasonal forecast of rainfall, but less is known about the impact of ENSO on the Benguela Current west of southern Africa. There is a significant weak correlation between ENSO and the Benguela Current upwelling sea surface temperature (SST) in austral summer. Correlation is positive for southern Benguela and negative for northern Benguela. A significant correlation exists with up to 8 months lag when ENSO leads. The impact of ENSO is due to weaker-than-normal upwelling favorable southeasterly winds during El Nino in southern Benguela, leading to warmer-than-normal coastal SST. In contrast, during La Nina, stronger-than-normal southeasterly winds lead to cooler-than-normal SST. The opposite effect applies to northern Benguela. The coastal wind change is part of an ENSO large-scale basinwide perturbation in the tropical and South Atlantic. However, non-ENSO-related SST variation in the Benguela upwelling can be as important as ENSO-related SST perturbation, and some ENSO events do not lead to the expected changes. Changes in the Benguela upwelling are linked to changes in the intensity of the trade winds associated with a change of the South Atlantic anticyclone intensity and position. In southern Benguela, changes are also associated with variations in midlatitude low pressure systems and associated upwelling unfavorable westerly winds. La Ninas favor the development of Benguela Ninos in Angola and Namibia. This study shows the potential for SST seasonal predictability in the Benguela upwelling due to the leading lag correlation between ENSO and the Benguela upwelling SST.

Rainfall extremes are of major and increasing importance in semi-arid countries and their variability has strong implications for water resource and climate impacts on the local societies and environment. Here, we examine extremes intraseasonal descriptors (ISDs) in austral summer rainfall (November–February) over South Africa (SA). Using daily observations from 225 rain gauges, ERA5 reanalysis and satellite estimates (TRMM-3B42), we propose a novel typology of wet extreme events based on their spatial fraction, thus differentiating large- and small-scale extremes. Long-term variability of both types of extreme rainfall events is then extensively discussed. The relationship between these two types of rainfall extremes and different modes of climate variability is further explored at multiple timescales. At low-frequency modes, rainfall extremes are assessed at interannual (IV: 2−8 years) and quasi-decadal (QDV: 8−13 years) timescales which are primarily associated with El Niño Southern Oscillation (ENSO) and Interdecadal Pacific Oscillation (IPO) respectively. At high-frequency modes, rainfall extremes are evaluated with synoptic-scale variability related to seven convective regimes of Tropical Temperate Troughs (TTTs: 3–7 days) and intraseasonal variability associated with eight phases of the Madden-Julien Oscillation (MJO: 30–60 days). The results demonstrate that using 7% of spatial fraction simultaneously exceeding the local threshold of the 90th percentile produces remarkable results in characterizing rainfall extremes into large- and small-scale extremes. Austral summer total rainfall is found to be primarily shaped by large-scale extremes which constitute more than half of the rainfall amount under observation, and nearly half in ERA5. Observation (ERA5) shows an average of 8 ± 5 (20 ± 7) days per season associated with large-scale extremes, which are comprised in 5 ± 3 (10 ± 3) spells with an average persistence of at least 2 days. Overall, we find a strong dependence of total rainfall on the number of wet days and wet spells that are associated with large-scale extremes. We also find that large- and small-scale extremes are well-organized and spatially coherent yet extreme conditions during small-scale events are found sporadic over the region, contrasting with large-scale events for which extreme conditions are found over a larger and coherent region. Teleconnections with global SSTs confirm that La Niña conditions favor overall wet conditions and wet extremes in SA. The frequency of large-scale extremes is consistently related to warmer SSTs in the North Atlantic while their link with warmer Indian and tropical South Atlantic Ocean found stronger without ENSO influence. At low-frequency timescale, risk ratio assessment shows that the frequency (total rainfall) of large-scale extremes is significantly modified by IV (QDV) timescale. We note strong variations in the frequency (total rainfall) of large-scale (small-scale) extremes when IV timescale lies in strong positive phase (i.e., +0.5 standard deviation). At high-frequency timescale, the synoptic-scale variability associated with TTT events, are mostly responsible for changes in large-scale extremes as nearly 75% of such events occur during early to mature TTT regimes (3−5) whereas small-scale extremes were found equiprobable during all synoptic regimes. A risk ratio assessment suggests that the probability of large-scale extremes in TTT regime 5 significantly enhance (suppress) during MJO phases 6−8 (1−2).

Abstract. In this paper, we review observational and modelling results on the upwelling in the inner tropical Atlantic. We focus on the physical processes that drive the seasonal variability of surface cooling and upward nutrient flux required to explain the seasonality of primary productivity. We separately consider the equatorial upwelling system, the northern coastal upwelling system of the Gulf of Guinea and the tropical Angolan upwelling system. For the equatorial regime, we discuss the forcing of upwelling velocity and turbulent mixing as well as the underlying dynamics responsible for thermocline movements and current structure. The coastal upwelling system in the Gulf of Guinea is concentrated along northern boundary and is driven by both, local and remote forcing. The particular role of the Guinea Current, nonlinearity and the shape of the coastline are emphasized. For the tropical Angolan upwelling, we show that this system is not wind-driven, but instead results from the combined effect of coastally trapped waves, surface heat and freshwater fluxes, and turbulent mixing. Finally, we review recent changes in the upwelling systems associated with climate variability and global warming and address possible responses of upwelling systems in future scenarios.

The seasonal cycle of sea surface salinity (SSS) along the Angolan coast is investigated using observations and a regional ocean model. The model reproduces the main characteristic of the seasonal cycle of SSS along the Angolan coast, such as the freshwater discharge signature off the Congo River plume and the low-salinity observed in February/March and October/November along the Angolan coast. The model also reproduces the two maxima of salinity in June/July and December/January. The analysis of the model salt budget reveals that the semi-annual cycle of SSS is controlled by the meridional advection of surface water, the vertical advection of subsurface water, and the mixing at the base of the mixed layer. The meridional advection is controlled by the Angola Current which brings low-salinity water from offshore region of the Congolese coast toward the south Angolan coast in February/March and October/November. The vertical advection contribution is modulated by the vertical stratification of salinity and not by vertical velocities which peak during the main Angolan upwelling season. The vertical stratification is due to the low-salinity intrusion at the Angolan coast that creates a strong vertical salinity gradient with low-salinity at the surface and high salinity at the subsurface.

This study examines sea level change in the context of decadal-scale variability in the ocean-atmosphere dynamics of the tropical Atlantic. This time scale is of great significance for adaptation and mitigation in the context of increasing societal threats from the ongoing harmful effects of anthropogenic climate change. Decadal climate variability in the Atlantic is caused by the interaction of the gyres and is evidenced by persistent multi-year anomalies in sea surface temperature, heat content and thermocline depth (through steric sea level and dynamic height). In this study, the tropical sea level anomaly (SLA) was decomposed into interannual and decadal time scales via an empirical orthogonal function (EOF) method. The SLA variability was investigated and found to be closely related to climatic variability patterns. In addition, decadal SLA variabilities were observed between 1993 and 2016, with SLA and SLP seasonal shifts occurring in the second decade, with no change in the equatorial wind stress, responsible for warm events.

The Benguela Upwelling System is one of the four most productive fisheries areas in the world, and it is therefore important to understand the mechanisms leading to changes at the decadal scales before developing scenarios or forecasts for the future of the region. In this study we examine, the changing characteristics of the Benguela upwelling system SST at the decadal scale and their links with the large-scale climate signal. We focused on Southern Benguela. Wavelet analysis shows three significant time scales of variability over the twentieth century: interannual (2-8 years), quasi-decadal (9-14 years) and interdecadal (19-26 years). The correlations between Southern Benguela SST and SST over the global ocean are different at both quasi-decadal and interdecadal time scales. At the interdecadal scale, the tropical and subtropical oceans, especially in the Pacific Ocean appear to be strongly linked to Southern Benguela SST fluctuations. The correlation between the tropical Pacific (ENSO-like) and Pacific Decadal Oscillation (PDO) SST and Southern Benguela SST are statistically significant. At the quasi-decadal time scale, the Southern Benguela SST is linked to the whole South Atlantic Ocean. The correlation chart between Southern Benguela SST anomalies filtered at the 9-14 years and the South Atlantic SST anomalies is reminiscent of the South Atlantic Subtropical Dipole mode (SASD), the dominant mode of variability in the South Atlantic Basin.

Regional sea levels often behave significantly differently from the global average, making it difficult to establish future sea-level projections. In the Atlantic Ocean, the regional average steric (thermosteric, halosteric) sea level plays an important role in the variability of the overall trend, associated with heat and freshwater, redistribution due to circulation, and freshwater input from melting land ice and river runoff over the past two decades. This contribution varies in space and time. Based on sea level measurements obtained by satellite altimetry from CMEMS products and salinity and temperature data from Argo floats for the period 2005-2015, we found that the Gulf of Guinea and the Atlantic Niño boxes experienced a large thermosteric relative to the Amazon box, which experienced a larger halosteric contribution to sea-level change. This remarkably large halosteric contribution is associated with a cooling in the upper 700 m range. Currently, local atmospheric forcing, such as wind variability, may not explain this warming while the Tropical Northern Atlantic (TNA) index tends to explain the freshening.