Magma‐Assisted Continental Rifting: The Broadly Rifted Zone in SW Ethiopia, East Africa

The Gofa Province and Chew Bahir Basin in the Broadly Rifted Zone (BRZ) between the southern Main Ethiopian Rift (sMER) and the northern Kenya Rift (nKR) record early volcanism and associated faulting in East Africa; however, the spatiotemporal relationships between volcanism and faulting remain poorly constrained. We applied apatite (U‐Th)/He (AHe) and zircon (U‐Th)/He (ZHe) thermochronometry to Neoproterozoic basement rocks from exhumed footwall blocks of the extensional Gofa Province and Chew Bahir Basin, and analyzed our result in the context of well‐dated regional volcanic units in the BRZ to unravel the interplay between tectonic exhumation, faulting and volcanism. Single‐grain AHe ages ranging from 1.0 to 136.8 Ma were recorded in 32 samples, and single‐grain ZHe ages from three samples range between 142.2 and 335.6 Ma. The youngest AHe ages were obtained from the Chew Bahir Basin and the narrow deformation zone in the Gofa Province. Our thermal modeling results reflect little or no significant regional crustal cooling prior to extensive volcanism, which started at about 45 Ma. Conversely, new and previously published thermal history models suggest that widespread crustal cooling related to regional extension occurred between ∼27 and 20 Ma. Thermal modeling results from subsets of samples indicate that following this initial diffuse extensional deformation, renewed exhumation occurred along a narrow zone within the Gofa Province and the Chew Bahir Basin during the middle to late Miocene (15‐6 Ma) and Pliocene (<5 Ma), respectively. The crustal cooling phases follow a regional trend in volcanic episodes. For example, initial cooling between 27 and 20 Ma corresponds with the end of widespread flood‐basalt volcanism (45–28 Ma), suggesting that spatially diffuse normal faulting may have initiated shortly after the emplacement of voluminous and areally extensive flood basalts. The Miocene and Pliocene shifts in deformation along the Mali‐Dancha and Bala‐Kela basins in the Gofa Province and the Chew Bahir Basin, respectively, may indicate strain localization during the late stage of rifting and ongoing tectonic interaction between the sMER and the nKR. Our results support the notion of crustal weakening by massive volcanism as a precursor to widespread extensional faulting, and thus offer further insights into magma‐assisted deformation processes in the East African Rift System.

• Initial diffuse normal faulting across the Broadly Rifted Zone of southern Ethiopia occurred soon after the emplacement of flood basalts during the late Oligocene • Extensional deformation migrated toward a narrow zone along the Mali-Dancha and Bala-Kela basins in the Gofa Province during late-stage rifting • Thermal modeling results reveal a close temporal relationship between exhumation and regional pulses in magmatism, indicating magma-assisted continental rifting

Supporting Information:
Supporting Information may be found in the online version of this article. 10.1029/2022TC007651 2 of 26 toward the interior of these tectonic depressions, and a narrow, volcanically and tectonically active rift develops during more advanced stages of extension (e.g., Corti, 2008;Ebinger et al., 1993;Morton et al., 1979;Richter et al., 2021;Riedl et al., 2022).
Two-dimensional numerical modeling studies indicate that upwelling of hot asthenospheric material may cause a reduction in lithospheric thickness, ultimately leading to crustal weakening and thus promoting extension as observed in the East African Rift System (EARS) (e.g., Buck, 2006).Efficient crustal extension can be further facilitated by the reactivation of inherited basement heterogeneities, such as suture zones, deep-seated crustal-scale shear zones, metamorphic fabrics, and structures associated with paleo-rifts (Corti, 2008;Kendall & Bertelloni, 2016;Vauchez et al., 1997).
Typically, large continental rift zones evolve from isolated tectonic basins during the early stages of extension into larger, throughgoing, and topographically linked depressions (e.g., Ebinger & Scholz, 2012;Ebinger et al., 1999;and references therein).Before this connectivity between individual basins is achieved, kinematic linking between these basins is accommodated within structurally complex transfer zones by the generation of closely spaced normal faults, the formation of local transfer faults, and pronounced magmatic activity accompanied by diking (Bosworth & Morley, 1994;Morley et al., 1992;Riedl et al., 2022).
The BRZ comprises multiple, N-S to NE-SW-oriented parallel rift basins between the sMER and the nKR (Figure 1).In Ethiopia, these areas include the Omo Valley and the Beto, Baneta, Mali-Dancha, Chew Bahir, Segen, Gelana, and Chamo basins (Figure 1b).These basins are filled with Tertiary and Quaternary sediments and volcanics, whereas the rift flanks are mainly comprised by Neoproterozoic basement rocks and overlying Eo-Oligocene flood basalts (Figures 1b and 2, e.g., Davidson, 1983;Steiner et al., 2022).
Concerning the temporal patterns of rifting in the sMER, low-temperature thermochronological studies at the uplifted margins of the Chew Bahir and the Beto (Gofa Province) basins in southern Ethiopia indicate that rifting and tectonic exhumation initiated at ∼20 Ma (Pik et al., 2008) and ∼12 Ma (Balestrieri et al., 2016;Philippon et al., 2014), respectively.Although several authors have proposed propagation of rifting from northern Kenya toward the Gofa Province via the Chew Bahir Basin in Ethiopia (Bonini et al., 2005;Chorowicz, 2005;WoldeGabriel et al., 1990), Balestrieri et al. (2016) suggested that synchronous deformation along the basin-bounding faults of the Beto Basin and the eastern margin of the Amaro Horst occurred between ∼12 and 10 Ma.Boone et al. (2019) also suggested synchronous rift initiation across the BRZ, but with an onset between 20 and 17 Ma (Figure 1b).A proper assessment of the timing of rifting in the sMER is further complicated by shifts in the locus of tectonic activity.For example, tectonic activity in the northern Gofa Province in the late Miocene to early Pliocene has subsequently shifted toward the currently seismically active Chew Bahir Basin (Ebinger et al., 2000).Furthermore, whereas WoldeGabriel and Aronson (1987) considered the Gofa Province as a failed rift segment and the Chew Bahir Basin in the south as an early-stage rift, Erbello et al. (2022) recently demonstrated the existence of active Quaternary normal faults, young tectonic landforms, and tectonically forced drainage networks in the northern Gofa Province along the Bala-Kela Basin margin.
From these different views, it can be concluded that the spatiotemporal characteristics of rifting in the Chew Bahir Basin and the Gofa Province are ambiguous and still poorly constrained.In addition, despite previous thermochronological studies and the availability of geochronological data, potential relationships between clustered  2019) and this study.The indicated apparent displacement was estimated from the TanDEM-X DEM (12-m resolution), and all the profile sections were vertically exaggerated by 5:1.
volcanic and tectonic activity in the sMER remain unclear.In this study, we therefore attempt to clarify these ambiguities and hypothesize that pulses in volcanism are intimately linked with phases of tectonic activity.Using new low-temperature thermochronological data in combination with a synopsis of radiometrically dated volcanic rocks and existing thermochronological data, we examine the spatiotemporal relationships between volcanism, rifting, and the evolution of topography in the BRZ.

Tectonic Setting
The BRZ is a structurally complex extensional deformation zone with imprints of multiple tectonic events.NW-SE-striking tectonic lineaments inherited from Cretaceous-Paleogene paleo-rifts obliquely cut NNE-SSW-striking Neoproterozoic basement fabrics (Bosworth & Morley, 1994;Brune et al., 2017;Emishaw & Abdelsalam, 2019;Morley et al., 1992; Figure 1a).Whereas the well-developed N-S to NE-SW-striking Cenozoic rift structures are closely aligned with the regional basement fabrics (Corti, 2008;Kendall & Bertelloni, 2016), the Cretaceous-Paleogene NW-SE-striking lineaments are associated with the Melut and Muglade basins in Sudan and South Sudan, and the Anza Rift in northern Kenya (e.g., Bosworth & Morley, 1994).These rift basins are connected in the subsurface; however, the N-S to NE-SW-trending rift basins associated with the active EARS obscure these features (e.g., Ebinger et al., 1993;Emishaw & Abdelsalam, 2019;Mamo, 2012).Recently, Kounoudis et al. (2021) imaged a NW-SE-trending area of high seismic-wave speeds at a depth of <200 km, suggesting that the pervasive band of crustal heterogeneity may have controlled rift localization in the BRZ during the Mesozoic and Cenozoic.

Geologic Setting
Neoproterozoic basement rocks are the oldest exposed units in the region; they consist of high-grade amphibolite and layered granulites that are exposed across the entire width of the BRZ (Gichile, 1992).The spatially extensive basement rocks were metamorphosed during Pan-African collisional orogenic processes between 750 and 550 Ma (Asrat & Barbey, 2003) and are characterized by NNE-SSW-striking foliations (Davidson, 1983).The granulites are mainly exposed along the southwestern margin of the Chamo Basin, on the Konso Plateau (Kazmin et al., 1978), and northeast of the Chew Bahir Basin (Davidson, 1983) (Figure 1b).The outcrops in the Konso Plateau are characterized by E-W-striking foliations (Asrat & Barbey, 2003;Davidson, 1983), reflecting effects of post-Pan-African tectonic events (de Wit & Chewaka, 1981).
Well-indurated, thin basal conglomeratic sandstone beds unconformably overlie the crystalline basement rocks (Davidson & Rex, 1980;Ebinger et al., 1993).A 5 to 30-m-thick conglomeratic sandstone layer, silicified at the top, is exposed along the margin of the Amaro Horst (Ebinger et al., 1993;WoldeGabriel et al., 1991) but is absent west of the BRZ (Davidson, 1983;Philippon et al., 2014).The age of the sandstone layer is uncertain; however, several authors have correlated the sediments with the petrographically similar lower unit of the Turkana Grits of inferred Cretaceous age in northern Kenya (Levitte et al., 1974;Murray-Hughes, 1933;Owusu Agyemang et al., 2019;Savage & Williamson, 1978;Walsh & Dodson, 1969).In contrast, Davidson (1983) suggested an early Paleogene depositional age, based on sediment composition and the nature of contact with the overlying Eo-Oligocene volcanics.The undeformed conglomeratic sandstone layer lacks organic remains and is thought to represent sedimentation across a peneplain surface prior to the emplacement of the voluminous Eo-Oligocene volcanics (Davidson, 1983;Davidson & Rex, 1980).
The thickness of the Amaro Basalt increases toward the north along the Amaro Horst, where the full sequence is exposed (Levitte et al., 1974; Figure 1b).To the south of the Amaro Horst, petrographically similar basalts directly overlying the crystalline basement rocks yielded a K-Ar age of 42.5 ± 0.7 Ma (e.g., WoldeGabriel et al., 1991).
To the west of the Amaro Horst, along the margins of the Chamo and Beto basins, the Gamo Basalts are overlain by the widespread Amaro Tuff, dated between 37.9 and 35.1 Ma (Davidson, 1983;Ebinger et al., 1993).Rooney (2017) correlated the Gamo basalts with the Mekonnen basalts (35-28 Ma; Davidson, 1983), and reclassified them as the Gamo-Mekonnen basalts; these volcanic rocks represent the main phase of flood-basalt volcanism between ca.38 and 28 Ma (Steiner et al., 2022).
During the early to middle Pliocene, thin lava flows of the Gombe Group were emplaced along the Omo-Turkana Depression (Watkins, 1986).The lava flows extend farther north toward the Usno Basin in the lower Omo Valley and crop out in the form of erosional remnants along the western margin of the Chew Bahir Basin (Davidson, 1983).
Based on petrographic analysis and K-Ar dating, Haileab et al. (2004) correlated these units with the Gombe Group basalts along the northeastern sectors of the Lake Turkana Basin (Watkins, 1986).These widely distributed basalt flows erupted during a short period between 4.05 and 4.18 Ma (Erbello & Kidane, 2018).
During the late Pliocene, volcanic activity shifted toward the southeast of the BRZ along the Ririba Rift segment (Corti et al., 2019).Outcrops of Quaternary volcanics are restricted to the northern sector of the Chamo Basin and to areas west of the Omo Valley (Figure 1b).The Quaternary Bobem (0.66 ± 0.02 Ma) and Nech Sar (0.99 ± 0.10 Ma) basalts in the northern Chamo Basin cap southwest-dipping fluvio-lacustrine deposits (Ebinger et al., 1993).Both units subsequently cut by N-S-striking normal faults (Ebinger et al., 1993).N-S-aligned Quaternary volcanic fields of the Kurath range in the Omo Valley were emplaced at ∼0.09 Ma (Jicha & Brown, 2014).
With respect to the rift-related magmatic evolution of the study area, it can be concluded that the volcanic rocks in the BRZ were emplaced in three major eruption periods during the Eo-Oligocene, the Miocene, and the Pliocene-Holocene.These distinct eruptive episodes were virtually coeval with phases of tectonic activity in the Red Sea Rift (e.g., Bosworth & Stockli, 2016), the eastern Tanzanian Craton (e.g., Rooney, 2020), and the nKR (Muirhead et al., 2022;Riedl et al., 2022).

Methods
Low-temperature apatite and zircon (U-Th-Sm)/He (AHe and ZHe) thermochronology is based on the decay of U, Th, and Sm isotopes, producing α-particles retained below system-specific closure temperatures that vary with cooling rate (Dodson, 1973), grain size (Farley, 2000;Reiners et al., 2002), and the amount of radiation damage in a crystal (e.g., Flowers et al., 2009;Gautheron et al., 2009;Guenthner et al., 2013).Temperature ranges in which α-particles are only partially retained are referred to as partial retention zones (PRZ).The AHePRZ and ZHePRZ range from ∼40 to 120°C and ∼40-220°C (Ault et al., 2019 ;Gautheron et al., 2009;Guenthner et al., 2013; and references therein), respectively.Assuming a mean surface temperature of 20°C and a geothermal gradient of 40°C/km (Pik et al., 2003), the range of thermal sensitivity of AHe and ZHe in this part of Africa corresponds to a depth of about 0.5-5.0km.Hence, these thermochronologic systems can potentially document the thermal history associated with tectonic and erosional processes in the uppermost part of the crust.

Sample Collection and Treatment
Thirty-five bedrock samples were collected from exhumed amphibolite and granulite rocks exposed in footwall blocks adjacent to the Chew Bahir (samples C1-12), Weyito (W1-2), Mali-Dancha (M1-9), Baneta (B1-4), Bala-Kela (K1-6), and Chamo (G1-2) basins (Figure 1b).As the extent of basement exposures diminishes laterally north of the Chew Bahir Basin (Davidson, 1983), we collected samples from both elevation profiles and along-strike of the exposed footwall blocks (Figures 1b and 2), similar to the approach chosen by Mortimer et al. (2016).Rock samples from six elevation profiles were collected from regions with sufficiently uplifted footwall blocks.A set of samples from the NW-SE-striking Mali-Dancha Basin margin was collected between 880 and 1,500 m asl.Two elevation profiles (C7-9 and C10-12), spaced ∼10 km apart, were obtained along a transect across the northeastern margin of the Chew Bahir Basin.A single sample (C-6) was also collected from the base of the exposed footwall farther south (Figure 1b).Additionally, two samples from different elevations were collected from transects across the flanks of the Baneta (B2-3), Mali-Dancha (M7-8), and Bala-Kela (K1-2) basins (Figures 1b and 2), respectively.At locations where exhumed Neoproterozoic basement rocks were characterized by insufficient relief for age-elevation profile sampling (<250 m) (Figure 2), single samples were collected from the base of the exposed footwall margin.
Following standard magnetic and heavy-liquid mineral separation, a total of 182 apatite and 10 zircon single grains were carefully handpicked, their dimensions measured, and packed in Pt (for apatite) and Nb (for zircon) tubes for isotope analyses at the University of Potsdam (He) and the GFZ-German Research Centre for Geosciences (U, Th, Sm), according to methods described in Zhou et al. (2017) and Galetto et al. (2021).We included isotope-specific stopping distance variations to obtain alpha-ejection (FT) corrected AHe and ZHe ages, following methods described in Ketcham et al. (2011) and Hourigan et al. (2005), respectively.

Thermal History Modeling
Thermal history modeling was performed using both the QTQt (Gallagher, 2012) and HeFTy (Ketcham, 2005) inversion programs.QTQt employs a Bayesian trans-dimensional Markov Chain Monte Carlo algorithm to search time-temperature paths that are compatible with observed data based on a posterior probability, whereas HeFTy employs a Frequentist approach using a random, non-learning Monte Carlo time-temperature search algorithm to determine thermal histories that agree with observed data based on a goodness-of-fit statistical test.The contrasting modeling and visualization outputs from QTQt and HeFTy each have their own unique advantages (cf., Abbey et al., 2023;Gallagher & Ketcham, 2018;Murray et al., 2022;Vermeesch & Tian, 2014), which we strategically exploited through a systematic and iterative modeling approach to thoroughly assess and interpret our (U-Th)/He data set.
To account for the impact of radiation damage on He diffusivities, the accumulation of radiation damage and annealing models of Flowers et al. (2009) and Guenthner et al. (2013) were applied in all QTQt and HeFTy inversions for the AHe and ZHe systems, respectively.Single-grain AHe data from each sample were first scrutinized for outliers by assessing potential age-eU-grain size relationships.Individual samples revealing distinctive age-eU-grain size relationships within all grains over a comparatively broad range of eU concentrations (>10 ppm) were first identified, and data from these samples were considered robust inputs for thermal modeling.
Potential outliers based on first-order age-eU relationships from each sample were first noted, and model sensitivities to the inclusion or exclusion of such grains were evaluated.Potential outliers were identified and excluded if their inclusion as inputs (a) significantly obscured the resulting thermal history in QTQt leading to poor age predictions for all grains and/or (b) prevented HeFTy from finding any "good" time-temperature history paths after ∼100,000 iterations.All excluded data points are identified in Table S1 of Supporting Information S1.
Thermal histories for selected individual samples were first modeled in QTQt, followed by multiple-sample modeling of compiled data from elevation profiles.All modeled thermal histories initiated at high temperature (450 ± 50°C) in the late Neoproterozoic to early Cambrian (525 ± 25 Ma) to represent the Pan-African orogenic event (Asrat & Barbey, 2003;Davidson, 1983;Yibas et al., 2002).Because the stratigraphic thickness for the Paleogene sediments and the Eo-Oligocene volcanics on top of the basement rocks is so variable, with the units even absent in some places (e.g., Davidson, 1983;Ebinger et al., 2000;Philippon et al., 2014), a broad range of temperature constraints extending from 120 to 20°C and covering a time interval between 80 and 30 Ma was applied.For elevation-profile samples, a regionally estimated geothermal gradient of 40°C/km (Pik et al., 2003) was used.A final constraint of 20 ± 10°C at 0 Ma, simulating present-day conditions, was applied to all models.
After implementing these time-temperature constraints, all models were allowed to search for solutions between 550 Ma and the present.We used 50,000 burn-in and 200,000 post-burn-in sampling iterations for individual inverse modeling experiments.The resulting expected thermal history model was constructed from all possible time-temperature paths weighted by their posterior probabilities (cf., Gallagher, 2012).A summary of thermal modeling input parameters and the model setup is presented in Table S1 of Supporting Information S1.
The probability distribution from the expected output of the time-temperature models from QTQt was used to set up the thermal history inversions with HeFTy.Samples that contained single-grain AHe data spanning a range of eU concentrations >10 ppm, with a comparatively distinctive age-eU relationship were selected for thermal inversions in HeFTy.The motivation to additionally model these data in HeFTy was twofold: (a) to confirm the thermal histories produced in QTQt for robust interpretation, and (b) to permit the forward modeling of each individual and reasonably fitting thermal history in order to visually compare predicted age-eU patterns to observed data (e.g., Guenthner, 2021).For each HeFTy inversion, an identical constraint box to that used in QTQt was imposed from 80-30 Ma to 120-20°C, followed by a second overlapping constraint from 70 to 1 Ma and 125-20°C to allow the model to explore potential post-80-Ma reheating and cooling scenarios.Each HeFTy inversion was also unconstrained during the pre-80-Ma time window to permit the exploration of varying radiation-damage accumulation histories that may have a direct influence on observed AHe data (e.g., Colleps et al., 2021), and thus impact the post-80-Ma thermal history.The details of each QTQt and HeFTy inversion, including justifications for geological constraints and data inputs, are presented in Table S1 of Supporting Information S1.Following each HeFTy inversion, the model output was exported to forward model age-eU and grain-size trends for each "good" t-T pathway using a modified Matlab script from Guenthner (2021), and the forward model predictions were compared to the observed data.
For each QTQt modeling result, we estimated the onset of rapid cooling and its uncertainty (earliest and latest possible onset of rapid cooling relative to each other; e.g., Abbey et al., 2023;Murray et al., 2022) by objectively defining inflection points (i.e., rapid cooling start times, Figure 3e) from the 95% confidence envelope (e.g., Balestrieri et al., 2016).When either of the inflection points was poorly expressed, defining the earliest or the latest possible onset of rapid cooling was aided by the corresponding relative probability distribution plot from the QTQt modeling results.The cooling rate for each model was estimated from the slope of the expected time-temperature path between specified time intervals (Table 3).

Apatite and Zircon (U-Th)/He Ages
Where possible, two to 12 single grains ranging in size from 38 to 145 μm (equivalent spherical radius-ESR) were analyzed from each sample in an attempt to quantify intra-sample age dispersion.Unfortunately, only one grain could be analyzed from samples B-1, B-4, C-3, and C-12, due to their poor apatite yield and quality.A summary of the data is presented in Tables 1 and 2. The position of the thermochronological samples and their structural context are shown in the geological cross sections in Figure 2.
Single-grain ZHe ages measured from sample C-11, collected at an intermediate structural position from the northeastern ridge flanking the Chew Bahir Basin (Figure 1b), range between 142.2 ± 1.9 Ma and 260.6 ± 3.0 Ma.Additional ZHe ages measured from samples C-1 and W-2, collected from the southwestern flank of the Chew Bahir and Weyito basins, respectively, range between 248.5 ± 3.9 Ma and 335.6 ± 4.5 Ma (Table 2).
The intra-sample age-dispersion (standard deviation of single-grain ages divided by sample mean) in our data set is large, with values ranging from ∼4 to 158%.Only a few of the analyzed samples (C-1, C-10, K-1, M-3, W-1, and W-2) exhibit dispersion within 20% of the mean age.Individual sample plots depicting measured age against ESR and eU concentration are presented in Supporting Information S1.The measured ages plotted against eU or ESR exhibit a weak relationship in most samples; however, some samples obtained from the foot of the Bala-Kela (K-1, K-6), Baneta (B-3), and Mali-Dancha (M-1, M-6) basin margins show a moderate to strong age-eU relationship.Sample M-6 shows a moderately positive age-eU trend, but its eU ranges only between ∼3 and 5 ppm (see Supporting Information S1).1T is one termination.2T is two termination).Table 1 Continued

Thermal History Modeling Results
Thermal history modeling was carried out for individual samples and sets of elevation-profile samples from the region.Here we present the mean possible onset of rapid cooling estimated from the average of the earliest and latest possible rapid cooling start times for the selected models (for further details see Table 3 and Table S3 in Supporting Information S1).Thermal modeling results from laterally offset elevation-profile samples C7-9 and C10-12 (with C-11 including ZHe dates) collected from the northeastern flank of the Chew Bahir Basin (Figure 1b) record a mean possible onset of protracted slow cooling (∼0.7-1.2°C/Myr) between ∼55 and 41 Ma (Figures 3a, 4a, and Figure S3 in Supporting Information S1).Similar slow cooling throughout the Cenozoic is recorded by samples C-3 and W-1, which were collected from the base of the southwestern Chew Bahir Basin margin and the NW-SE-striking Weyito Basin flank, respectively (Figure S3 in Supporting Information S1).
Spatially distributed samples from the Chew Bahir Basin and the Gofa Province exhibit an onset of cooling (>∼2°C/Myr) between 27 and 21 Ma.For example, individually modeled samples M-1 and M-2, collected from different elevations across the NW-SE-striking Mali-Dancha Basin flank, record an onset of rapid cooling (>∼2.3°C/Myr) between ∼25 and 14 Ma (Figures 3b and 4a).When modeled jointly, samples M-1 and M-2 exhibit episodic cooling with onset of rapid cooling at ∼27 and 4 Ma (Figure 3b).Farther southeast along the margin of the Mali-Dancha Basin, sample C-5 from the transition zone between the Mali-Dancha and Chew Bahir basins (Figures 1b and 4a) records well-constrained, protracted cooling after ∼23 Ma (Figure 3a).Additionally, samples M-9 and K-6, collected from the transition zone between the Mali-Dancha and Bala-Kela basins and the eastern flank of the Bala-Kela basin (Figures 1b and 4a), respectively, exhibit an onset of cooling between ∼23 and 21 Ma (Figures S3b and S3d in Supporting Information S1).
Samples B-2 and B-3, collected from different elevations across the NE-SW-striking Baneta Basin margin (Figures 2 and 4a), document rapid cooling starting between ∼15 and 12 Ma (Figure 3c).A consistent onset of rapid cooling (∼5°C/Myr) at ∼15 Ma was recorded when jointly modeling samples B-2 and B-3 (Figure 3c).In the Gofa Province, sample K-1 collected from the base of the Bala-Kela basin (Figures 2 and 4a) records a positive correlation between age and grain size (Figure S2 in Supporting Information S1) and exhibits slow cooling through the AHePRZ prior to more rapid cooling (∼5°C/Myr) commencing at ∼6 Ma (Figure 3d).
In summary, thermal history modeling results for samples from the base of the rift flanks in the BRZ indicate an onset of cooling in time intervals between 55 and 40 Ma, 27-21 Ma, and 15-6 Ma (Figures 4a and 5).More recent (<6 Ma) rapid cooling (∼6°C/Myr) is documented from sample C-2, which was collected at the base of the Chew Bahir Basin margin (Figures 3a and 4a), and elevation-profile samples M-1 and M-2 from the Mali-Dancha Basin margin ( Figures 3a, 3b, and 4, and Table 3).

Table 2 Zircon (U-Th)/He Data Summary
Most of the HeFTy modeling results for the selected samples (Figure S3 in Supporting Information S1) predict onset times of cooling that are similar to the corresponding QTQt modeling predictions.For example, HeFTy modeling results for samples M-2, B-3, and K-1 reflect a consistent onset of rapid cooling with the QTQt modeling results (Figure 3).Additionally, HeFTy and QTQt modeling results for sample C-2 exhibit a similar timing for the onset of rapid cooling (Figure 3a), suggesting that the QTQt modeling results for the selected samples are robust.
The predicted ages from thermal histories obtained from QTQt modeling results fit the observed data reasonably well.Multiple single-grain ages from samples M-1, M-2, B-3, K-1, and K-6 were predicted within <∼20% error of the observed ages.In contrast, some of the predicted ages from samples C-2, K-1, M-1, and M-2 differ from the observed ages by ∼30-50% (Figure 3a).The misfit between predicted and observed ages from the jointly modeled samples C7-9, M1-2, and B2-3 ranges between 5% and 50%, with few misfits exceeding 50% of the observed ages (Figure 3).Except for two single-grain ages, all predicted ages for the elevation-profile samples C10-12 lie within ∼20% of the observed ages (Figure S3a in Supporting Information S1).The age-eU patterns predicted from individual "good" thermal history outputs for each sample modeled by HeFTy demonstrate that the modeling results obtained are robust despite the dispersion in the observed data (Figure 3).While there are many factors that can influence the reproducibility of AHe ages (including zoning, radiation damage, chemistry; e.g., Flowers et al., 2023), the fact that the sampled bedrock is Neoproterozoic and was subjected to multiple tectono-thermal events prior to rift-related extension and final exhumation (Bosworth & Morley, 1994;Chorowicz, 2005;Ebinger et al., 2000;Morley et al., 1992) implies that the analyzed apatite grains must have experienced a complex thermal history that may have amplified the variable effects of damage accumulation, zonation, and grain chemistry on AHe data disparities.

History of Vertical Movements Prior to the Emplacement of Flood Basalts
Thermal history modeling results of our new AHe and ZHe data from the BRZ record little or insignificant crustal cooling shortly prior to massive volcanism in southern Ethiopia between 45 and 28 Ma (Davidson, 1983;Ebinger et al., 2000).Only the time-temperature histories obtained from samples located mainly along and across Note.p Jointly modeled elevation profile samples.Onset of cooling was determined based on the shape of the 95% credible interval (inflection point), marking change in cooling rate.The cooling rate was also calculated from the slope of the weighted mean expected thermal history between a specified time interval.stdv Standard deviation of the mean onset of cooling.

Table 3 Thermal Modeling Results for Individual and Set of Elevation Profile Samples
the Chew Bahir Basin flanks (Figure 3a) and from the base of the Weyito Basin margin (Figure S3 in Supporting Information S1), characterized by comparatively older AHe ages, document slow cooling, likely recording denudation, since ∼55 to 40 Ma (Figure 4).For example, the elevation-transect samples from the northeastern footwall block of the Chew Bahir Basin (C7-9, C10-12) (Figure 5b) reveal onset of slow cooling between 50 and 40 Ma (Figures 3a and 5).Likewise, sample W-1 from the base of the NW-SE-striking Weyito Basin margin (Figure 5b) records protracted, slow cooling throughout the Cenozoic (Figure S2b in Supporting Information S1).A similar prolonged residence close to the surface and slow cooling from shallow crustal levels can be inferred from the analysis of sample C-3 from the base of the southwestern margin of the Chew Bahir Basin (Figures 3a and 4a).
Denudation associated with this early onset of slow cooling may result from Paleogene extension related to the opening of the Anza Rift in northern Kenya (Bosworth & Morley, 1994;Morley et al., 1990Morley et al., , 1992) ) or regional uplift documented in the Ogaden Basin in the east (Mége et al., 2015), and ensuing erosion.Farther to the northeast, a sample from the Segen Basin margin (Figure 1b) with a central AFT age of ∼65 Ma (Balestrieri et al., 2016) and a mean AHe age of ∼61 Ma (Boone et al., 2019) exhibits a consistent thermal history of prolonged residence close to the surface during the Cenozoic (Boone et al., 2019).This thermal history may further indicate that crustal cooling was associated with the attainment of NW-SE-trending local relief during the early Paleogene, assuming that adjacent regions were low-relief surfaces during this period (Davidson & Rex, 1980).(Balestrieri et al., 2016;Boone et al., 2019;Pik et al., 2008, coding as in Figure 1b); samples from this study are indicated by their sample ID (Table 1).(b) Cooling rate calculated from the slope of expected tT-path in specified time intervals.(c) Inferred lineaments and earthquakes (from the International Seismological Centre-www.isc.ac.uk) in the sMER and the Chew Bahir Basin.
Based on thermal modeling results from a set of elevation-profile samples from the Beto Basin margin and AFT cooling ages ranging between 75 and 60 Ma, Philippon et al. (2014) inferred a period of rapid crustal cooling during the Cretaceous and Paleogene due to regional uplift documented in northern Kenya (Boone et al., 2019;D. A. Foster & Gleadow, 1996;Spiegel et al., 2007;Torres Acosta et al., 2015).Combining the AFT data from Philippon et al. (2014) with new AHe ages, Boone et al. ( 2019) also inferred rapid crustal cooling (∼2.5-5°C/Myr) across the NW-SE-striking flank of the Beto Basin during the Cretaceous.Rapid cooling may have been linked to the formation of the Anza Rift and associated NW-SE-oriented structures that form pronounced lineaments that appear to be present-day vestiges of these early extensional processes in the BRZ (Bosworth & Morley, 1994;Emishaw & Abdelsalam, 2019;Kounoudis et al., 2021;Morley et al., 1992).This interpretation is compatible with a fault-kinematic analysis of structures recorded in the Turkana Grits of northern Kenya that documents NE-SW-oriented extension during the Paleogene (Vetel & Le Gall, 2006).Flank uplift associated with activity in the Anza Rift might have generated the source rock for the petrographically correlated and up to 30-m-thick Cretaceous-early Paleogene (?) conglomeratic sandstones exposed in the southern sMER (WoldeGabriel et al., 1991) and in the Turkana area of northern Kenya (Levitte et al., 1974).

Implication for Interaction Between the Southern Main Ethiopian and Northern Kenya Rifts
Rapid Plio-Pleistocene cooling along the footwall flanks of the Chew Bahir and the Mali-Dancha basins reflects the current tectonic interaction between the sMER and nKR (Figure 3a and Figure S3 in Supporting Information S1: C-2 and M1-2); this interaction is likely to have involved a reactivated, NW-SE-striking inherited basement fabric via the seismically active Segen Basin (Figures 4c and 6).The reactivated basement fabrics and the structurally influenced course of the Segen River (Figures 4c and 7d) suggest that the current hydrologic connectivity between lakes Abaya, Chamo, and Turkana via the Chew Bahir Basin might have been forced by Pleistocene tectonic processes.This interpretation is supported by (a) the aborted propagation of the sMER toward the Ririba Rift during the early Pleistocene (e.g., Corti et al., 2019), (b) the current active tectonism of the southern Gofa Province and the Chew Bahir Basin (Erbello et al., 2022;Philippon et al., 2014), as well as (c) observed seismicity clusters along the Omo Valley, the southern Gofa Province, and the Chew Bahir and Segen basins (Musila et al., 2022).Finally, geodetic survey results (i.e., Knappe et al., 2020), documenting high strain  (Davidson, 1983;Davidson & Rex, 1980;Ebinger et al., 1993Ebinger et al., , 2000;;Franceschini et al., 2020;George et al., 1998;George & Rogers, 2002;Haileab et al., 2004;Jicha & Brown, 2014;Rooney, 2010)  rate of ∼4.5 mm/yr within a narrow zone between the Chew Bahir and Segen basins (Figure 6), and a high strain further support the inferred tectonic interaction between the sMER and the nKR (Figure 7e).
For example, the early Paleocene to late Eocene thermal history with slow cooling (0.7-1.2°C/Myr) documented by a set of elevation profiles from the northeastern margin of the Chew Bahir Basin (C7-9 and C10-12), and the Pliocene rapid crustal cooling recorded by elevation-profile samples M1-2 and sample, C-2 collected from the NW-SE-striking Mali-Dancha Basin margin and the western flank of the Chew Bahir Basin, respectively, furthermore strengthen the interpretation of local recent onset of crustal cooling in the southern Gofa Province and the Chew Bahir Basin (Figure 5).In light of these observations, recent seismicity (Musila et al., 2022), historical earthquake records (https://www.isc.ac.uk), and structural data (Figure 4c), we suggest that the current, and historical kinematic interaction between the sMER and the Chew Bahir Basin (Ebinger et al., 2000;Erbello et al., 2022;Philippon et al., 2014) began during the early Pliocene, following reactivation of NW-SE-striking inherited basement fabrics (Figures 4c and 7d).

Rifting and Magmatism
Our new thermochronological results, combined with published records from previous studies, allow discussing the overall spatiotemporal characteristics of exhumation, rift-basin development, and the timing of magmatism during the evolution of the BRZ.The onset of widespread crustal cooling throughout the BRZ corresponds with the end of ubiquitous flood-basalt volcanism between 45 and 28 Ma (Davidson, 1983;Ebinger et al., 2000;Rooney, 2020;Steiner et al., 2022).Whereas it is possible that widespread post-magmatic crustal cooling may reflect regional thermal relaxation following the end of the Eo-Oligocene volcanic phase (e.g., Braun, 2016) or effects related to local magmatic diking during the Miocene (e.g., Murray et al., 2018), we argue, for reasons discussed below, that the documented crustal cooling most likely reflects exhumation facilitated by areally extensive faulting that may have begun shortly after emplacement of the basalt flows (Figures 5, 7b, and 7c).
The flood basalts in southwestern and northern Ethiopia were emplaced in three phases: (a) 49-46 Ma (Akobo Basalts); (b) 42-33 Ma (Main Series); and (c) 35-28 Ma (Mekonnen Basalt) (e.g., Davidson, 1983;Steiner et al., 2022).In southwestern Ethiopia, a principal, spatially extensive eruptive phase that generated the Amaro and Gamo-Mekonnen magmatic units (Steiner et al., 2022) occurred during a prolonged period between 45 and 28 Ma.After this Eo-Oligocene magmatic phase, ∼500-m-thick basaltic flows were emplaced during the Miocene, between ∼18 and 11 Ma (e.g., Bonini et al., 2005;Davidson, 1983;Ebinger et al., 1993Ebinger et al., , 2000)).Eruptive centers for the Miocene basalts are mainly located along the eastern margin of the Gidole Plateau with a few volcanic centers mapped in the northeastern part of the Weyito Basin (Ebinger et al., 1993), indicating minor magmatic diking activity throughout the Chew Bahir Basin and the Gofa Province during the Miocene (Ebinger et al., 2000).During the Pliocene and the Quaternary, volcanic activity continued along the axial zone of the sMER, in the Omo Valley, southeast of the Chew Bahir Basin, and in the Ririba Rift (Bonini et al., 2005;Corti et al., 2019;Davidson, 1983;Ebinger et al., 1993Ebinger et al., , 2000;;Franceschini et al., 2020;WoldeGabriel et al., 1991), indicating prolonged regional magmatic activity (Rooney, 2010).Based on geophysical data from the nKR, Wheildon et al. (1994) documented an increase in the geothermal gradient from the rift margin toward the currently active axial zone of the rift, reflecting a steady increase in geothermal gradient during rifting (e.g., Morgan, 1984).Considering the prolonged magmatic activity during rifting (Rooney, 2010) and the associated increase in the geothermal gradient over time in the study region (Wheildon et al., 1994), it is problematic to consider regional thermal relaxation to have occurred between the Miocene and Pliocene.Furthermore, the inferred minor magmatic diking activity in the investigated region (Ebinger et al., 2000) suggests that the documented onset of widespread crustal cooling across the BRZ is more likely linked to regional exhumation associated with diffuse faulting and not to effects associated with local magmatic activity.The widespread diffuse faulting, however, would have been assisted by thermal weakening of the entire lithosphere through magmatic activity.This interpretation is consistent with previous studies suggesting that extensional deformation in the Chew Bahir Basin and the Gofa Province is mainly accommodated by slip along the border faults (e.g., Bonini et al., 2005;Boone et al., 2019;Ebinger et al., 2000;Philippon et al., 2014).
Using geochemical characteristics, Steiner et al. (2022) showed that the main phase of Eo-Oligocene volcanism was pronounced due to the significant volumes of magma that must have moved through the magmatic plumbing system compared to prior magmatic phases.Such massive volcanism suggests lithospheric weakening through magmatic diking (e.g., Buck, 2006), which thermally preconditioned the lithosphere for subsequent stretching (Steiner et al., 2022).In addition to these thermal processes, the mechanical strength of rocks in this region and farther south in Kenya may have been further reduced by inherited crustal anisotropies associated with NNE-SSW-striking Neoproterozoic basement fabrics that facilitated rift-parallel fracture propagation and normal faulting (e.g., Bosworth & Morley, 1994;Corti et al., 2013;Ebinger et al., 2000;Morley et al., 1992;Philippon et al., 2014) as well as transfer faulting between different basement-foliation domains (Hetzel & Strecker, 1994;Philippon et al., 2014).
Following the phase of diffuse, regionally widespread faulting, deformation progressively became more localized along the narrow zone of the Mali-Dancha and Bala-Kela basin margins in the Gofa Province and propagated northward during the late stages of rifting (Figures 4a and 7d); this propagation of rifting was accompanied by resumed regional volcanism between 18 and 12 Ma (Davidson, 1983;Ebinger et al., 2000), suggesting a close tectono-magmatic relationship (Figure 5).
The time-temperature history results of some samples (e.g., M1-2, B-2, and M-9) analyzed in this study exhibit a progressive increase in temperature between the Eocene and early Miocene (Figure 3 and Figure S4 in Supporting Information S1).Given the overall regional emplacement of >1.5-km-thick basaltic flows (i.e., the Amaro, Gamo, and Makonnen basalts) between ∼45 and 28 Ma (Davidson, 1983), it is reasonable to expect that some of our sampled rocks might have undergone reheating during this period.A coeval reheating event documented from the adjacent flanks of the Beto and Gelana basins (Balestrieri et al., 2016;Boone et al., 2019) is interpreted to constrain the regional extent of the emplaced volcanic products during the Eocene and Oligocene.
Overall, the thermal modeling results obtained from spatially distributed samples along and across the footwall blocks of the Chew Bahir Basin and Gofa Province suggest synchronous, tectonically driven crustal cooling.For example, time-temperature histories from samples located across the NW-SE-oriented flank of the Mali-Dancha (M1-2) and from along the Bala-Kela (K-6) basins appear to record the onset of cooling at rates of >2°C/Myr between 27 and 21 Ma (Figures 3b  and 3d), supporting a previously inferred regional onset of faulting at ∼20 ± 2 Ma (Pik et al., 2008).Sample M-9, which was collected from the base of the rift flank at the transition zone between the Mali-Dancha and Bala-Kela basins, records a consistent trend of protracted rapid cooling (∼3.1°C/Myr) that initiated at ∼23 Ma (Figure S2b in Supporting Information S1).Moreover, the synchronous onset of cooling across the BRZ followed spatially extensive regional volcanism between 45 and 28 Ma (Steiner et al., 2022), and predates the onset of divergent Nubia-Somalia plate motion at ∼20 Ma (DeMets & Merkouriev, 2016;Iaffaldano et al., 2014), suggesting that crustal weakening via magmatic diking (e.g., Buck, 2006;Kendall & Bertelloni, 2016)-possibly accompanied by the effects of mantle-drag forces (e.g., Brune et al., 2023;Ebinger et al., 2000)-may have triggered synchronous faulting across the BRZ between ∼27 and 20 Ma (Table 3 and Figure 5).
As shown above, following the spatially diffuse faulting in the BRZ, deformation progressively became more localized within a narrow zone and propagated toward the northern Gofa Province over time (Figures 4a and 5).Modeling results obtained from samples B-2 and B-3, collected at different structural positions across the NE-SW-striking Baneta basin margin, suggest rapid cooling to have started at ∼15 Ma, with a rate of ∼5°C/Myr (Figure 3c).These results may indicate that extensional faulting occurred in the Baneta halfgraben area during the middle Miocene.The geological cross sections constructed across the rift in the BRZ reveal a similar basinward migration of faulting over time (Figure 2).The tectonically related crustal cooling during the formation of the Baneta halfgraben at Ma and the down-faulting of the Baneta halfgraben (Figure 2) thus suggests basinward strain localization during the middle Miocene (Figures 4a and 5a), a phenomenon that has also been observed in other extensional regions during advanced rifting (e.g., Agostini et al., 2011;Baker et al., 1988;Melnick et al., 2012;Morton et al., 1979).Similarly, in the Gelana and Chamo basins faulting has evolved from more diffuse spatial patterns toward locally focused extension in the center of the basins at ∼13 Ma and during the Quaternary, respectively (Balestrieri et al., 2016;Boone et al., 2019;Ebinger et al., 1993;Levitte et al., 1974).
The basinward migration of deformation in the southern Gofa Province propagated farther toward the Bala-Kela Basin in the late Miocene at ∼6 Ma.To the west of the Bala-Kela Basin, consistent spatiotemporal patterns regarding the onset of faulting between the southwestern (20-17 Ma) and northwestern (8-6 Ma) Beto Basin margin have been reported by Boone et al. (2019) and demonstrate temporal variation in the onset of faulting between the southern and northern Gofa Province.
The discontinuous onset of faulting in the southern and northern parts of the Gofa Province suggests that lateral fault propagation stalled during the growth of the rift basins.Fault-propagation patterns in such settings may be modified and influenced by many factors, including stratigraphic layer thickening (Benedicto et al., 2003), rheological contrasts (Roche et al., 2012), and the effects of inherited deformation fabrics (Camanni et al., 2019;Corti, 2008;Molnar et al., 2019).In particular, the presence of pre-existing deformation fabrics oriented at a high angle with respect to the developing rift may inhibit fault propagation (e.g., Molnar et al., 2019).Using the results of a seismic tomographic study, Kounoudis et al. (2021) imaged a NW-SE-trending pervasive band characterized by high seismic-wave speeds below the southern Gofa Province and the northern Chew Bahir Basin.The authors concluded that the ∼40-km-wide zone has a sharp, nearly vertical boundary that consistently follows the trend of structures inferred to be related to the Anza Rift.This region also coincides with an increase in surface topography (>1.5 km) attributed to northward thickening of the Eo-Oligocene flood basalts (Davidson, 1983), possibly suggesting a lithospheric strength contrast between the southern and northern sectors of the Gofa Province.The presence of these inherited heterogeneities between the different basins may have significantly influenced the northward propagation of deformation and hence resulted in spatial disparity in the onset of faulting.Thermal history modeling results from the southern and northern Gofa Province (B2-3 and K-1, respectively) exhibit this disparity in the onset of cooling (Figures 4a and 5a and Table 3).The nature of crustal cooling in these areas is interpreted to reflect the Cenozoic extensional reactivation of these anisotropies.Therefore, we propose that the spatial disparity in the onset of rapid cooling between the southern and northern sectors of the Gofa Province may explain the current basin architecture, that is, a wider and deeper basin in the south resulting from earlier rifting and a narrow basin in the north related to a later onset of rifting.A northward decrease in fault displacement (Figure 2) and Quaternary fault-scarp characteristics that indicate ongoing tectonic activity along the Bala-Kela margin of the northern Gofa Province (Erbello et al., 2022) is consistent with the recent onset of rapid cooling along the western margin of the Bala-Kela Basin (Figures 4 and 5).
The basinward shift in deformation and northward propagation of faulting during late-stage rifting temporally corresponds with a renewed pulse in volcanism.Phonolite eruptive centers aligned parallel to the orientation of the Mali-Dancha and Bala-Kela basins were active between 16 and 12 Ma (Davidson, 1983).Farther to the east, along strike of the Gidole margin (Figures 4 and 5), eruptive centers of the Getera-Kela basalts dated between 18 and 12 Ma (Ebinger et al., 2000;George & Rogers, 2002) suggest that faulting may have been accompanied by magmatism during late-stage rifting; however, strain accommodation in that region was mainly through slip along the border faults during this period (Ebinger et al., 2000).Pliocene-Holocene volcanic activity following the orientation of the Omo Valley (Jicha & Brown, 2014), the southeastern margin of the Chew Bahir Basin (Ebinger et al., 2000), and the sMER (Corti et al., 2013;Philippon et al., 2014) furthermore emphasize the importance of magmatism during rifting within the BRZ.For regions farther south, in the southern Turkana Rift, Rooney et al. (2022) suggested migration of deformation toward a narrow zone accompanied by spatially focused magmatism.Similarly, a recent structural and geochronological study in the nKR documented a significant influence of magmatism associated with Quaternary strain localization within the inner graben of the rift (Riedl et al., 2022), similar to observations made in the central and southern sectors of the Kenya Rift (Baker, 1986;Baker et al., 1988;Strecker et al., 1990).
Taken together, our new thermochronological data and modeling results recording diffuse initial faulting between 27 and 20 Ma are consistent with earlier studies that inferred synchronous faulting within the BRZ at ∼20 ± 2 Ma (Boone et al., 2019;Pik et al., 2008), and the migration of deformation toward the northern Gofa Province over time (Balestrieri et al., 2016;Bonini et al., 2005;Boone et al., 2019;Philippon et al., 2014).Furthermore, our data suggest that initial diffuse faulting followed the end of massive volcanism between ∼45 and 28 Ma (Figure 5).Moreover, the documented crustal cooling associated with faulting in the Chew Bahir Basin and the Gofa Province reveals a close temporal relationship with regional pulses in magmatism (Figure 5a), and thus supports the notion of magma-assisted continental rifting suggested elsewhere in the EARs (e.g., Rooney, 2020).

Conclusions
The combination of low-temperature thermochronological apatite and zircon (U-Th)/He data, and thermal modeling results in southern Ethiopia, as well as published regional 40 Ar/ 39 Ar data and seismicity data, provide insight into the tectono-geomorphic evolution of the Chew Bahir Basin and the Gofa Province from the Paleogene to the present (Figure 7).Additionally, our study helps to characterize the nature of the tectonic interaction between the sMER and nKR.The main outcomes of the study are summarized below: 1.The early Paleocene to late Eocene exhumation associated with slow cooling in the BRZ may have been linked to the opening of the Anza Rift in northern Kenya, suggesting it was the source area for the spatially restricted Paleogene Turkana Grits in northern Kenya and compositionally similar conglomeratic sandstones in the sMER.2. Spatially diffuse faulting associated with tectonically controlled rapid crustal cooling that occurred between 27 and 20 Ma corresponds with the end of ubiquitous flood-basalt volcanism, which was active between 45 and 28 Ma (Steiner et al., 2022).This temporal relation implies that faulting initiated after the emplacement of flood basalts, supporting the notion of crustal weakening through thermal processes during active rifting.Thermal modeling of samples collected from across the BRZ reveals this close temporal relationship between exhumation and regional pulses in magmatism, indicating magma-assisted continental rifting.3. The migration of deformation along a narrow zone of the Gofa Province between ∼15 and 9 Ma corresponds to a renewed pulse in volcanism (18-12 Ma), suggesting a progressive change in deformation mechanisms from tectonic faulting to magma-assisted rifting.The propagation of faulting toward the northern Gofa Province at ∼6 Ma reflects a temporal north-to-south variation in rift-basin development, likely due to inherited crustal heterogeneities.4. Tectonic interaction between the sMER and nKR began during the early Pliocene and was facilitated by reactivation of the NW-SE-striking zones of weakness in the basement that are inferred to have formed during Cretaceous extensional processes associated with the evolution of the Anza Rift.

Figure 2 .
Figure 2. AHe cooling ages superposed on across-rift geological cross sections (A-A′; B-B′; C-C′; locations shown in Figure 1b).AHe cooling ages are from Pik et al. (2008), Boone et al. (2019) and this study.The indicated apparent displacement was estimated from the TanDEM-X DEM (12-m resolution), and all the profile sections were vertically exaggerated by 5:1.

Figure 3 .
Figure 3. Thermal history results obtained from QTQt(Gallagher, 2012), and HeFTy(Ketcham, 2005) modeling for individual and multiple samples.The thermal modeling results are grouped by region: (a) Chew Bahir; (b) Mali-Dancha; (c) Baneta; and (d) Bala-Kela basins, and a thermal model example (e) modified fromGallagher (2012).The expected thermal history (solid black line) represents the weighted mean of all the individual time-temperature history paths.For thermal history models for elevation-profile samples, the topmost sample is plotted in blue and the lowermost sample is plotted in red.The thermal histories for intervening samples are shown in gray.For individual and elevation-profile samples, the 95% credible interval is drawn as a stippled line.The relative probability is color-coded, for legend see panel (a).The black boxes are the thermal history constraints.Only "good" fit thermal history paths are indicated for the HeFTy modeling results.For individually modeled samples (C-2, M-1, M-2, B-2, B-3, K-1 and K-6), HeFTy modeling results of the "good" fit thermal history paths (gray lines) and forward modeling of the thermal history paths (blue) are shown beside each QTQt models in order to visually compare predicted age-eU patterns to the observed data (orange).The light-pink vertical band indicates a range of rapid cooling start times.

Figure 4 .
Figure 4. Thermal modeling results of selected samples from the BRZ on shaded relief maps.(a) Onset of cooling.The transparent blue envelope represents a narrow zone of deformation during the late stage (<10 Ma) of rifting.Circles with letters in them denote published data from the Chew Bahir, Beto, and Galana basins(Balestrieri et al., 2016;Boone et al., 2019;Pik et al., 2008, coding as in Figure1b); samples from this study are indicated by their sample ID (Table1).(b) Cooling rate calculated from the slope of expected tT-path in specified time intervals.(c) Inferred lineaments and earthquakes (from the International Seismological Centre-www.isc.ac.uk) in the sMER and the Chew Bahir Basin.

Figure 6 .
Figure 6.W-E-oriented topographic swath profile (80 km width) showing GPS-derived eastward velocities (1σ error bars) (modified from Knappe et al. (2020)).Note the 1 mm/yr difference in velocity west and east of Segen Basin, indicating active extension in the region.

Figure 7 .
Figure 7. Conceptual tectonic-geomorphic evolution of the BRZ: (a) spatially restricted, NW-SE-trending local relief developed during the Late Cretaceous and prior to ∼45 Ma; (b) prolonged massive volcanism (45-28 Ma; Steiner et al., 2022); (c) diffuse faulting and isolated rift-basin development (CB, Chew Bahir; MD, Mali-Dancha; Be, Beto; W, Weyito; Ba, Baneta; BK, Bala-Kela; Se, Segen; C, Chamo; A, Abaya; and Ge, Gelana) after the end of flood-basalt emplacement; (d) migration of deformation toward a narrow zone of the rift and northward fault propagation; (e) strain localization in the southern Gofa Province and Chew Bahir Basin, and tectonic interaction between the Chew Bahir Basin and the sMER.The solid black arrows indicate regional kinematic changes from the Cretaceous (∼NE-SW) to the present (∼E-W).The black arrow denotes extension direction.