Paleozoic-Mesozoic thermal evolution along the East European
Platform margin based on kerogen thermal maturity analysis
combined with apatite and zircon low temperature
thermochronology in NE Poland

Abstract. The Phanerozoic tectono-thermal evolution of the SW slope of the East European Platform (EEP) in Poland is reconstructed by means of thermal maturity, low temperature thermochronometry and thermal modelling. We provide a set of new thermochronometric data and integrate stratigraphic and thermal maturity information to constrain the burial and thermal history of sediments. Apatite fission track analysis (AFT) and zircon (U-Th)/He (ZHe) thermochronology have been carried out on samples of sandstones, bentonites, diabase and crystalline basement rocks collected from 17 boreholes located in central and NE Poland. They penetrated sedimentary cover of the EEP subdivided from the north to south into the Baltic, Podlasie and Lublin Basins. The average ZHe ages from Proterozoic basement rocks as well as Ordovician to Silurian bentonites and Cambrian to lower Carboniferous sandstones range from 848 ± 81 Ma to 255 ± 22 Ma with a single early Permian age of 288 Ma, corresponding to cooling after a thermal event. The remaining ZHe ages represent partial reset or source ages. The AFT ages of samples are dispersed in the range of 235.8 ± 17.3 (Middle Triassic) to 42.1 ± 11.1 (Paleogene) providing a record of Mesozoic and Cenozoic cooling. The highest frequency of the AFT ages is in the Jurassic and Early Cretaceous prior to Alpine basin inversion. Thermal maturity results are consistent with the SW-ward increase of the Palaeozoic and Mesozoic sediments thickness. An important break in a thermal maturity profile exists across the base Permian-Mesozoic unconformity. Thermal modelling showed that significant heating of Ediacaran to Carboniferous sedimentary successions occurred before the Permian with maximum paleotemperatures in the earliest and latest Carboniferous for Baltic-Podlasie and Lublin Basins, respectively. The results obtained suggest an important role of early Carboniferous uplift and exhumation at the SW margin of the EEP. The SW slope of the latter was afterward overridden in the Lublin Basin by the Variscan orogenic wedge. Its tectonic loading interrupted Carboniferous uplift and caused resumption of sedimentation in the late Viséan. Consequently, a thermal history of the Lublin Basin is different from that in the Podlasie and Baltic Basins, but similar to other sections of the Variscan foreland, characterised by maximum burial at the end of Carboniferous. The Mesozoic thermal history was characterised by gradual cooling from peak temperatures at the transition from Triassic to Jurassic due to decreasing heat flow. Burial caused maximum paleotemperatures in the SW part of the study area, where the EEP was covered by an extensive sedimentary pile. However, farther NE, due to low temperatures caused by shallow burial, the impact of fluids can be detected by VR, illite/smectite and thermochronological data.



25
The boundary between the old and thick East European Platform (EEP) and the younger and thinner Palaeozoic Platform of Western Europe is entirely concealed beneath Palaeozoic and Mesozoic basins filled with several kilometres thick sedimentary successions (Figs. 1-4;Guterch et al., 1986Guterch et al., , 1999Pharaoh, 1999;Grad et al., 2002;Mazur et al., 2015). These basins form an extensive platform cover resting upon the SW slope of the East European Craton, comprising Paleoproterozoic to Mesoproterozoic crystalline basement. The tectonic and thermal evolution of the area was punctuated by several successive 30 phases of extension and shortening related to the Caledonian and Variscan orogenies, early Permian continental rifting and Late Cretaceous-Paleocene basin inversion. The timing and scale of these processes have not been yet properly constrained owing to the scarcity of thermochronological studies performed in the area and incomplete sedimentary record on the SW margin of the EEP that is due to the intervening periods of exhumation (Fig. 2). Although quantitative estimates of exhumation have been previously performed based on thermal maturity of organic matter, they provided divergent results due to the absence of 35 thermochronological data (e.g., Botor et al., 2002;Poprawa et al., 2010). However, in the Scandinavian part of the EEP, the discordant zircon U-Pb ages, onlap of Palaeozoic sediments, geomorphological analyses and low-temperature thermochronology indicate that the Precambrian basement was buried beneath extensive Palaeozoic and Mesozoic sedimentary successions (Lidmar-Bergström, 1993Hansen, 1995;Larson and Tullborg, 1998;Larson et al., 1999;Hendricks and Redfield, 2005;Söderlund et al., 2005;Green and Duddy, 2006;Larson et al., 2006;Hendriks et al., 2007;Japsen et al., 2015Japsen et al., , 2018Guenthner 40 et al., 2017).
In this study, we reconstructed the Phanerozoic tectono-thermal evolution of the south-western margin of the East European Platform in Poland by means of thermal maturity, low temperature thermochronometry and thermal modelling. We provide a set of new thermochronometric data and integrate stratigraphic and thermal maturity information to constrain the burial and thermal history of sediments. Using apatite fission-track (AFT) and zircon (U-Th)/He (ZHe) analyses, we reconstructed   Figure 3. Simplified chronostratigraphic logs for the Podlasie-Lublin and Baltic Basins with the location of samples indicated (based on various sources includingŻelichowski, 1987; Modliński et al., 2010;Poprawa, 2010) . to the TTZ ( Fig. 4; Kanev et al., 1994;Nehring-Lefeld et al., 1997;Swadowska and Sikorska, 1998;Grotek, 1999Grotek, , 2006Grotek, , 2016Skręt and Fabiańska, 2009;Więcław et al., 2010Więcław et al., , 2012. In all stratigraphic units, thermal maturity increases toward SW correspondingly to the increase of burial depth. Accordingly, mean random vitrinite reflectance (VR) for the base of Phanerozoic sequences in the Baltic, Podlasie and Lublin Basins attains 5.0, 1.3 and 3.4% VR, respectively (Grotek, 1999,   VR -in the adjacent depth level in which VR was measured in shale (Nehring-Lefeld et al., 1997;Grotek, 1999Grotek, , 2005Grotek, , 2006Grotek, , 2016 An acceptable fit corresponds to thermal histories representing the t-T paths that give a goodness of fit (GOF) value greater than 0.05 for both the age and the length distribution (Ketcham, 2005). The minimum statistic value above 0.05 means that 180 all statistics pass the 95% confidence test. When the minimum is above 0.5, the statistical precision limit, the model is termed good. For a comprehensive overview of fission-track methods and their modelling techniques, see more details in Donelick (2005); Ketcham (2005) and Ketcham et al. (2007b). For the helium diffusion kinetics in zircon, we applied the model by Guenthner et al. (2013). Chronostratigraphic subdivisions and absolute ages refer to Gradstein et al. (2012) throughout the text. Palaeozoic strata in the NE part of the EEP and negligible or very low re-heating during Phanerozoic (Grotek, 1999(Grotek, , 2006(Grotek, , 2016Skręt and Fabiańska, 2009;Pletsch et al., 2010).
All other ZHe mean ages are significantly younger than a stratigraphic age and thus are considered partially or totally reset (Tab. 2, Fig. 1  Sample unweighted aver. ± 1 s.e.

He U Th Sm
Amount of helium is given in nano-cubic-cm in standard temperature and pressure. Amounts of radioactive elements are given in nanograms. Ejection correct. (Ft) -correction factor for alpha-ejection according to Farley et al. (1996) and Hourigan et al. (2005). Uncertainties of helium and the radioactive element contents are given as 1 sigma, in relative error %. Uncertainty of the single grain age is given as 2 sigma in Ma and it includes both the analytical uncertainty and the estimated uncertainty of the Ft. Uncertainty of the sample average ages are in 1 standard error, as (SD)/(n)1/2; where SD = standard deviation of the age replicates and n = number of age determinations; eU has been calculated with the formula eU = U + 0.235 · T h; eu -euhedral, i -inclusions, si -small inclusions, frac -fracture, elips -eliptic shape of grain , fi fluid inclusions. *320 ZHe age without

Apatite fission track data
The results of AFT analyses, performed on 21 samples, are presented in Table 3, Figure 1 and Appendix (Figs. A1 and A2).

225
A stratigraphic age of samples ranges from Proterozoic to Jurassic (Tab. 1). The quality of AFT data is varied, but in most samples as a maximum 20 apatite grains were studied (Tab. 3). The average uranium content is from 3 to 70 ppm. All the AFT ages represent unimodal age populations as shown by high P (χ 2 ), younger than their stratigraphic age except for sample

Thermal maturity
Thermal maturity data put key constraints on the interpretation of thermochronological results and ensuing thermal modelling.
Besides a countrywide compilation of thermal maturity data (Fig. 4), we discuss in detail profiles of key boreholes that penetrate the sedimentary cover of the East European Platform. We also refer to thermochronological results obtained on samples that were collected from these boreholes. Corrected present-day temperatures in Figures 6-12 are based on analysis of geophysical 250

Dosimeter Spontaneous Induced
Nc -number of measured crystals, ρs -density of spontaneous tracks (x 10 6 tracks for cm −2 ); Ns -number of counted spontaneous tracks; ρi -density of induced tracks in external detector (mica) (x 10 6 tracks for cm −2 ); Ni -number of counted induced tracks; ρd -density of induced tracks in external detector which cover dosimeter (glass CN5)(x 10 6 tracks for cm − 2); Nd -numbers of counted tracks. P (χ 2 ) [%] -probability homogeneity apatite population were showed by the test agreement χ 2 value (Galbraith, 1981). AFT age± 1 σ error is a central age of sample (Galbraith and Laslett, 1993) counted by using calibration method zeta (Hurford 1983(Hurford , 1990) and dosimeter (glass) CN5. Data analyses and age calculations were based on a Zeta value is 348.18 ± 6.52 (A. Anczkiewicz), and accomplished by using Trackkey 4.2 software (Dunkl, 2002). Crystals chosen for confined track measurements had a well-polished surface, parallel to the c-axis. U -amount of uranium was calculated by software Trackkey 4.2. MTL (µ m± SE) -mean confined length. No -number of measured confined tracks. SD -standard deviation. Dpar mean -mean etch pit diameter, in brackets are number of etch pit diameters measured. data available (Górecki et al., 2006a,b). Although, a number of AFT and ZHe samples per well with is usually low (in many wells just a single sample), but integration of thermochronological and thermal maturity data provides meaningful results.
One of the most prominent features of thermal evolution recognised in some boreholes is fluid flow overprint on earlier burial diagenesis effects. The influence of advective heat transfer might be inferred for any basin that displays following features (Ziagos and Blackwell, 1986;Middleton et al., 1994;Lampe et al., 2001;Green and Duddy, 2012): (1) palaeotemperatures being and scattered to be suitable for calculation of VR gradient. Therefore, paleogeothermal gradient of 20°C/km was assumed for the Opalino-2 well that is equivalent to the paleogradient estimated for theŻarnowiec IG-1 well. In several adjacent wells, the 270 discontinuity of VR profile between Silurian and Permian (base Permian unconformity) suggests the pre-Permian development of thermal maturity ( Fig. 6b; Botor et al. (2019a). In the nearbyŻarnowiec IG-1 borehole (Fig. 4), a paleogeothermal gradient during maximum burial (20°C/km) seems to be similar to a present-day value (18°C/km) as both are almost parallel (Fig. 6) after omitting effects of pressure retardation. A paleogeotermal gradient of 20°C/km applied to the Silurian-early Carboniferous succession in the Opalino-2 borehole results in c. 5 km of exhumation. Furthermore, a temperature of 60°C is consistent with limited sedimentary burial of sample B13 (e.g., Botor et al., 2019a).
No episodes of significant Mesozoic exhumation are documented by the Gołdap IG-1 well, whose profile is full of sedimentary gaps but did not reveal any erosional unconformity. A significant discrepancy between VReq-derived paleotemperature (c. 75°C) and illite/smectite-derived paleotemperature (160°C) exists in the nearby Bartoszyce IG-1 well (Fig. 4). A similar 285 discrepancy between the maximum palaeotemperatures evaluated from illite-smectite and biomarkers was detected for the EEP by Derkowski et al. (2020). This diagenetic pattern was interpreted as the result of short-lasting pulses of potassium-bearing hot fluids, effectively promoting illitization in porous rocks without altering the organic matter. Correspondingly, we suggest that the reset of the AFT age in sample B13 was achieved in temperatures close to the upper limit of the apatite partial annealing zone in the presence of fluids. This interpretation is supported by paleotemperature profile of the nearby Bartoszyce IG-1 well, 290 where a bell-shape paleotemperature profile suggests transient fluid flow in Triassic rocks (Fig. 7d)  The approach by Petersen at al. (2013) was used to recalculate reflectance measurements on vitrinite-like macerals from the lower Palaeozoic strata into a standard vitrinite reflectance scale. Since the Opalino-2 VR profile is too short to define gradient and estimate exhumation the adjacent wellŻarnowiec IG-1 VR was used. The profile clearly shows overpressure retardation of thermal maturity, which occurs within the mudstone/claystone dominated Silurian section in most wells in the western Baltic Basin. TheŻarnowiec IG-1 VR profile shows a break across the Silurian/Permian unconformity as a VR value is 0.5% in Permian sediments. (C) Paleotemperature profile of the Opalino-2 (green diamonds) andŻarnowiec IG-1 (blue open diamonds) wells. Yellow line represents present-day geothermal gradient based on the corrected bottom hole temperature data (Górecki et al., 2006a,b). Paleogeothermal gradient of 20°C/km in the Opalino-2 well, comparable to the present-day gradient and the paleogradient in theŻarnowiec IG-1 well, gives exhumation of c. 4-5 km similar to that in theŻarnowiec IG-1 well. that is characterised by a discontinuity of the VR profile. Since the latter is relatively short an assessment of paleogradient and exhumation (c. 3 km) is uncertain and seems to be overestimated compared to regional data (Botor et al., 2019a). Another discontinuity of the VR profile occurs at the base Permian unconformity, indicating a pre-Permian cooling event (Fig. 7c).

Polik IG-1
The Ordovician bentonite (sample B36) from the Polik IG-1 borehole (Tab. 1) yielded a well-constrained ZHe mean age of 300 288 ± 6 Ma (Tab. 2; Fig. 8a). The Ordovician bentonite horizons gave also an illite K-Ar ages of 347-343 Ma (Kowalska et al., 2019), showing approximately the time of maximum paleotemperature. Therefore, the ZHe age probably represents cooling after the early Carboniferous thermal peak. A linear VR profile (Fig. 8b) shows a discontinuity across a base Permian unconformity between Ludlow (Silurian; 2.0% VR) and Rotliegend (Permian; 0.8% VR). Also, illite/smectite profile shows such a break (Kowalska et al., 2019). The paleogeothermal gradient of 35°C/km was calculated for the late Palaeozoic, the 305 assessment implying exhumation of the Silurian-Devonian, and likely lowest Carboniferous strata by 5.6 km (Fig. 8c). This estimate agrees with the maturity modelling results by Botor et al. (2019a). The inferred Mesozoic temperature profile for the Polik IG-1 borehole suggests 700 m of exhumation experienced by the late Mesozoic strata and paleogeothermal gradient of 23°C/km, similar to the present-day gradient of 24°C/km (Fig. 8c). concluded that sample B45 was exposed to a maximum paleotemperature of about 120-130°C. A nonlinear VR profile of the Tlłszcz IG-1 borehole is similar to that in the Siedliska IG-1 well (Fig. 10b), suggesting a fluid flow event (e.g., Ziagos and Blackwell, 1986). There are no Carboniferous strata in the Tłuszcz IG-1 borehole, where a base Permian unconformity incises upper Silurian sediments (Fig. 9). At the bottom of the Permian-Mesozoic succession, paleotemperature was relatively high, reaching 90°C in Permian sediments. The illite/smectite data indicate 120°C at the top of the Silurian strata that is higher than 320 the VR-derived paleotemperature, the relationship indicative of possible fluid flow influence (Fig. 9c).

Siedliska IG-1
Bashkirian ( . succession in the area. There is a regional unconformity and stratigraphic gap between the upper Carboniferous (Moscovian) 325 and Middle Jurassic. A non-linear VR profile below the unconformity (Fig. 10) shows that fluid flow may have contributed Other explanations as in Figures 6, 7. to maturation of organic matter (e.g., Ziagos and Blackwell, 1986). A discontinuity in the vitrinite reflectance profile exists at the unconformity since thermal maturity of Mesozoic strata is below 0.5% VR (Fig. 10). Owing to a non-linear VR profile an assessment of post-Carboniferous exhumation might be overestimated, and it is not attempted herein.

Łopiennik IG-1 330
Sample B19 was collected from the bottom of the Viséan succession (Tab. 1, Fig. 5b) in the Łopiennik IG-1 borehole that is located in the Lublin Basin (Figs. 1, 4). Five zircons were dated in this sandstone sample by means of the ZHe method. An average ZHe age of this sample is 287 Ma after omitting the single grain age that is older than a stratigraphic age (562 Ma; Tab. 2, Fig. 11a). VR value of 1.0% for sampled depth shows a maximum paleotemperature of 140°C, suggesting only partial reset   . of the ZHe age (Fig. 11b). The Ediacaran to Carboniferous VR profile allows to estimate VR and paleogeothermal gradients 335 for the late Palaeozoic (Fig. 11b, c). Although no VR data exist in the upper part of Carboniferous and Mesozoic sections of the profile, average VR values in the adjacent wells are c. 0.8-1.0% and 0.45-0.55% VR for the uppermost Carboniferous and Jurassic-Cretaceous strata, respectively (Botor et al., 2002(Botor et al., , 2019aGrotek, 2005  for Palaeozoic (R 2 =0.80) allows to estimate post-Namurian exhumation at 3700 m (assuming 0.2% VR as an initial value).
Alternatively, it can be calculated from a paleogeothermal gradient (24°C; R 2 =0.89) that a 4300 m thick pile of the sediments was removed, assuming 20°C as a surface temperature from the late Carboniferous to Early Jurassic (Fig. 11c). Both estimates are similar to that from maturity modelling based on the VR and porosity data (Botor, 2018;Botor et al., 2019a). However, assuming overpressure retardation in Silurian sediments, a paleogeothermal gradient of 23°C would suggest exhumation up to 345 7 km (Fig. 11c).  the immature (below 0.65% VR) Mesozoic part (Fig. 15b). The Mesozoic and Palaeozoic VR profiles are shifted relative each 355 other across the unconformity. The VR profile break suggests significant exhumation between the Silurian and Middle Jurassic.
The LK1/1 sample, showing Permian ZHe ages, was heated up to 180°C according to the VR data (Tab. 1). However, the inferred Palaeozoic paleotemperature profiles of the Lubycza Królewska-1 and Narol IG-1/PIG-2 boreholes show unrealistic c. 11 km exhumation with a paleogeothermal gradient of c. 13°C/km. The measured VR profile is too short to confidently define a paleogeothermal gradient and, thus, the tentatively calculated 11 km exhumation is probably overestimated.

Thermal modelling
Samples yielding the best quality analytical results were chosen for thermal history modelling by means of the HeFTy software (Ketcham, 2005;Ketcham et al., 2007b).  Table 4. Except for the B47 sample (Jurassic sandstone, Tyniewicze IG-1) and B39 sample (Permian sandstone, Słupsk IG-1), the AFT and ZHe ages are considerably younger than the age of deposition. Therefore, we did not consider a pre-depositional thermal history of the detrital grains. The thermal modelling results are compiled in Figure 13, where the best fit results are shown. High goodnessof-fit (GOF) values (0.90 to 1.00) suggest that the resulted time-temperature (t-T) paths are plausible.

Baltic Basin
In the western part of the Baltic Basin, thermal modelling was performed for the O-2/2 sample (Upper Ordovician bentonite; Tab. 1-3, Fig. 1) from the Opalino-2 borehole using ZHe data only and based on the assumption of a positive correlation between effective uranium (eU = U + 0.235 · T h; in ppm) and ZHe single grain ages. Constrain boxes for the O2/2 sample were established to represent following regional subsidence or heating events: development of a Caledonian foreland basin  (360-330 Ma) that is represented in a sedimentary pile by erosional gap. This is consistent with the illite K-Ar ages (336-329 Ma) from the same area (Kowalska et al., 2019). Subsequent Variscan shortening had a little impact on the thermal model in accord with a tectonic setting of the Opalino-2 well that is located beyond the Variscan deformation front (Fig. 1). Therefore, the late Carboniferous-Mesozoic time slice appears the time of relative thermal stability although the best fit curve suggests 390 limited middle to late Mesozoic reheating. This corollary is difficult to verify owing to the lack of AFT data. Finally, the model shows acceleration of cooling coeval with the tectonic inversion of the Permian-Mesozoic basin.
In the eastern part of the Baltic Basin, the thermal maturity of lower Palaeozoic strata is one of the lowest in the Polish part of the EEP (Tab. 1; Fig. 4; Grotek, 2006Grotek, , 2016Pletsch et al., 2010). The sedimentary cover on the crystalline basement is the thinnest across the study area. Two samples were analysed in the Gołdap IG-1 borehole: Permian sandstone B13 and 395 Proterozoic granitoid B14 (Tab. 1, Fig. 1). Since a negative correlation between eU and ZHe single grain ages in the B14 sample did not allow for helium data modelling (Green and Duddy, 2018) only AFT data were used. Constrain boxes for sample B13 (earliest Jurassic). This temperature is higher than that predicted by VR and illite/smectite data (c. 60-70°C); (Kowalska et al., 2019). Therefore, we interpret early Mesozoic heating as a cumulative effect of increasing sedimentary burial and transient fluid flow in the late Permian and Triassic. The latter effect is confirmed by a bell-shaped paleotemperature profile in the nearby Bartoszyce IG-1 borehole (Fig. 7d). Consequently, the model predicts cooling after cessation of fluids influence for the remaining part of the Mesozoic and Cenozoic despite continued subsidence and sedimentation. Importantly, Mesozoic-  (Ketcham, 2005). The light green range corresponds to the envelope of thermal paths with acceptable fit (goodness of fit -GOF > 0.05), the magenta range shows the envelope for thermal paths with good fit (GOF > 0.5).
Bold dark blue curve shows weighted mean path, whereas thin black line is the best fit curve. Black rectangles correspond to constrain boxes.
Further explanations in the text. was the time of thermal stability with only minor acceleration of cooling at the time of the Permian-Mesozoic basin inversion (Fig. 13d).
In the Tłuszcz IG-1 borehole two samples were analysed. Thermal modelling for sample B44 (Lower Triassic sandstone; Fig. 13e) was performed using constrain boxes for: Triassic basin subsidence (250-230 Ma), further Mesozoic subsidence (220-150 Ma), and Permian-Mesozoic basin inversion (100-60 Ma). A thermal peak of 100°C is predicted by the weighted 445 mean path of the model for the earliest Jurassic (c. 200 Ma). This is close to a VR-based paleotemperature of 90°C in Permian sediments (Fig. 9). The post-Early Jurassic period was the time of gradual cooling despite ongoing sedimentation as already discussed above (model B13). The overlying sedimentary pile generated temperature of c. 50-55°C, using the Mesozoic geothermal gradient, that is less than 90°C noted in the Permian strata. A present-day temperature at a depth of sample B44 is only 36°C (Tab. 1). Therefore, the Mesozoic was the time of net cooling that was enhanced by 700 m of exhumation calculated 450 for the Permian-Mesozoic profile (Fig. 9).
Although the weighted mean path of the model does not show a clear temperature peak in the late Palaeozoic the best fit curve indicates a thermal peak of 110°C at 310 Ma (Fig. 13f). This prediction is in accord with a VR-derived paleotemperature of 116°C (Tab. 1). The thermal peak corresponds to a big erosional gap between the Silurian and Permian as well as a break in the paleotemperature profiles between the lower Palaeozoic and Permian-Mesozoic strata (Fig. 9). In the post-Carboniferous 460 period, the model shows cooling down to a present-day temperature of 55°C.

Lublin Basin
Thermal modelling was performed for upper Carboniferous (Bashkiria) sandstone sample B38 that was collected from the Siedliska IG-1 borehole (Figs. 1, 4). In the model for sample B38 (Fig. 13g), constrain boxes were set to represent a few succeeding subsidence and thermal events: Variscan shortening and subsequent continental rifting (320-260 Ma), subsidence 465 of the Permian-Mesozoic basin , two periods of further subsidence (180-150 and 145-110 M), and Permian-Mesozoic basin inversion (100-60 Ma). The model shows a temperature peak of 80°C (the weighted mean path) in the latest Carboniferous-earliest Permian. This is less than VR-derived paleotemperature in the range of 90-100°C (Fig. 10). However, the best fit curve reveals a maximum temperature of c. 110°C for roughly the same age range. The latest Carboniferousearliest Permian thermal peak is consistent with a big erosional gap between the Carboniferous and Jurassic and a break of 470 the paleotemperature profile (Fig. 10). The thermal peak was followed by gradual cooling through the Mesozoic and Cenozoic down to the present-day temperature of 35°C.
A positive correlation between e-U and ZHe ages in sample B19 of Carboniferous sandstone from the Łopiennik IG-1 borehole favours thermal modelling despite the lack of AFT data. Modelling was carried out for a single zircon grain B-19z4 due to dispersion of the ZHe ages (Tab. 2, Fig. 13h). Four constrain boxes were used in the model: 340-300 Ma (subsidence and subsequent shortening of the Variscan foreland), 280-240 Ma (initial subsidence of the Permian-Mesozoic basin), 230-150 Ma (further subsidence of the Permian-Mesozoic basin), and 100-60 Ma (Permian-Mesozoic basin inversion). The weighted mean path of the model reveals rapid heating until a thermal peak at 320 Ma. The best fit curve predicts a thermal peak in the early Permian (280 Ma). The results are consistent with a big erosional gap between the Carboniferous and Upper Jurassic as well as a major break in the paleotemperature profile across the unconformity (Fig. 11). The post-early Permian time was characterised 480 by a gradual cooling with some acceleration at the time of the Permian-Mesozoic basin inversion (Fig. 13h). For modelling of sample LK1/1 (Silurian bentonite; Fig. 13i), the following constrain boxes were assumed: (1)  Ma). The weighted mean path indicates a thermal peak of 140°C at c. 320 Ma. The best fit curve points to thermal peak of 485 180°C at c. 300 Ma. The latter modelled temperature is consistent with a VR-derived paleotemperature of 175°C (Fig. 12).
Furthermore, an age of c. 300 Ma is close to the illite K-Ar ages of 298-272 Ma from the same borehole (Kowalska et al., 2019). The latest Carboniferous thermal peak is consistent with a major erosional gap between Carboniferous and Jurassic in the Lubycza Królewska borehole and an important break of the paleotemperature profile across the unconformity (Fig.12).
The thermal peak is followed by gradual cooling through Mesozoic and Cenozoic down to the present-day temperature of 490 65°C (Fig. 13i).

Discussion
Characteristic features of the SW slope of the East European Platform in Poland is an increasing thickness of sediments towards the TTZ and an erosional unconformity at the top of the lower Palaeozoic. An exception is the Lublin Basin, where the unconformity is at the top of Devonian (e.g., Narkiewicz, 2007). Another regional unconformity is located at the base of 495 Permian to Jurassic strata. However, in the area, where Devonian and Carboniferous strata are missing (Fig. 14), there is only one major unconformity between the lower Palaeozoic and Permian. These unconformities must represent a succession of major tectonic events, but separation of their effects is impossible without thermal maturity and thermochronological data due to an incomplete sedimentary record.

Thermochronological constraints 500
The SW-ward increase of the total sediment thickness on the slope of the EEP has been well-known for decades based on borehole and seismic data (e.g., Młynarski, 1982;Poprawa and Pacześnia, 2002;Mikołajczak et al., 2019). Thermal maturity results are consistent with this trend (Fig. 4) since the highest paleotemperatures, based on VR data, are recorded in those boreholes that are located close to the TTZ (Tab. 1; Figs. 1, 4). Consequently, the only ZHe ages that may represent fully reset zircons are those obtained from the Polik IG-1 borehole (Tab. 2) in agreement with the VR-derived pre-Permian paleotemperature of 505 217°C (Fig. 8). An early Permian ZHe mean age of 288 Ma corresponds to cooling below the ZHe closure temperature of 130°C, the corollary consistent with the illite/smectite-derived paleotemperature (Fig. 8). The remaining ZHe ages represent partial reset or source ages (section 5.1). Therefore, the ZHe ages obtained, besides sample B36, cannot be directly used to constrain the time of cooling and must be supplemented by thermal modelling.
Considering significant thickness of the Permian-Mesozoic sedimentary cover on the SW slope of the EEP (Figs. 1, 4), it 510 is not unexpected that AFT ages are dispersed in the range of 235.8 ± 17.3 to 42.1 ± 11.1 providing a record of Mesozoic to Cenozoic cooling (section 5.2). The highest frequency of the AFT ages is in the Jurassic and Early Cretaceous (Tab. 3) i.e., they are older than inversion of the German-Polish Basin (e.g., Senglaub et al., 2005;Resak et al., 2010;Łuszczak et al., 2020).
Consequently, the results obtained suggest that cooling through the apatite partial annealing zone occurred in the Mesozoic before tectonic inversion in the Late Cretaceous. This might be related to the fact that tectonic inversion was significantly 515 weaker in the part of the German-Polish Basin onlapping the EEP (Krzywiec et al., 20017b). Although inversion structures are undoubtedly recognised in this area (Krzywiec, 2009) the offset of vertical movements is probably below the sensitivity of the AFT method. Further inferences on timing and rate of Mesozoic cooling can be derived from thermal modelling. Nevertheless, the AFT data are not suitable for constraining a pre-Permian thermal history of the area. Out of 9 models in total, 7 provides solutions for the part or entire Palaeozoic (Tab. 4, Fig. 13). Among the latter, two groups can be distinguished: (1) models predicting latest Devonian to early Carboniferous maximum paleotemperature followed by 525 rapid cooling throughout the Carboniferous and (2) models predicting a temperature peak at the end of Carboniferous ( 300 Ma), succeeded by cooling in the Permian-Mesozoic (Fig. 13). All 4 models in the first group were built for samples that were collected from boreholes in the Baltic and Podlasie Basins, the area missing Devonian and Carboniferous sedimentary rocks (Figs. 13,14). Despite a regional base Permian unconformity and an important shift of the VR profile across this unconformity (Figs. 6-9), the age of the unconformity and a Palaeozoic thermal event has remained so far unconstrained, being broadly 530 limited to the interval between the end of Silurian and beginning of Permian. Therefore, our models make an important step forward predicting maximum heating at the transition from the Devonian to Carboniferous and rapid cooling soon afterwards (Fig. 15). This result is consistent with the early Carboniferous illite K-Ar ages by (Kowalska et al., 2019).
The second group of models represents samples collected from boreholes in the Lublin Basin (Figs. 13, 14). Although in the Lubycza Królewska-1 borehole, Silurian strata are directly overlain by Jurassic sediments, in the Łopiennik IG-1 and 535 Siedliska IG-1 boreholes parts of Devonian and Carboniferous sedimentary sections are preserved (Figs. 10, 11). In the latter two, no break in the VR profile is observed across the boundary between the lower Palaeozoic and Devonian or Carboniferous (Figs. 10,11). This is consistent with the latest Carboniferous thermal peak modelled (Fig. 13)  is not associated with a major gap in the thermal history. In contrast to the Baltic and Podlasie Basins, a peak paleotemperature was achieved there in the latest Carboniferous.

Permian-Mesozoic thermal history
Two models for Permian and Lower Triassic samples show exclusively the Mesozoic thermal history (Fig. 13). They both reveal a paleotemperature peak at the transition from the Triassic to Jurassic and subsequent cooling (Fig. 13) Permian-Mesozoic basin (c. 1000-1500Figs. 7, 9) AFT and VR data suggest that burial heating was strengthen by transient fluid flow in the Triassic.

545
The Proterozoic granite sample (B14) from the Gołdap IG-1 borehole yielded the same Phanerozoic thermal history (Fig. 13c), including a peak temperature at the Triassic-Jurassic boundary and subsequent cooling.
Our data show that Mesozoic was the time of cooling despite ongoing sedimentation that might be considered a contradiction.
On the other hand, we know a present-day temperature at sampling depths and a maximum paleotemperature for sampled strata from vitrinite reflectance (VR) data (Tab. 1). Furthermore, VR data from are consistent with illite/smectite data, where 550 available. A comparison between VR data for the Permian-Mesozoic samples and present-day temperatures shows that sampled strata must have been cooled during Mesozoic-Cenozoic regardless the cause is. Of course, the simplest solution would be erosional unroofing. Indeed, the Permian-Mesozoic-Cenozoic sedimentary cover is relatively thin and full of sedimentary gaps. Nevertheless, the well logs did not show any clear erosional or ungular unconformities. Furthermore, paleogeothermal gradients calculated based on VR data suggest only little to no exhumation throughout the Mesozoic (Figs. 6-12). Therefore, 555 decreasing heat flow appears a main cause of Mesozoic cooling. Consequently, the mostly Jurassic-Early Cretaceous AFT data obtained, or earliest Jurassic thermal peaks predicted by thermal models do not represent any specific tectonic events or erosional episodes. Instead, they correspond to the time when the AFT samples left the apatite partial annealing zone. This is possible because the VR-derived paleotemperatures for the Permian-Mesozoic samples do not exceed 90°C (besides the Bodzanów IG-1 borehole with a present-day temperature of 75°C).

560
The models predicting an early Carboniferous thermal peak show relatively slow cooling during the Mesozoic (Fig. 13). In addition, the models built for sample O2/2 and B19 reveal acceleration of cooling at the time of Late Cretaceous basin inversion (Fig. 13a). The models characterised by a latest Carboniferous temperature maximum also demonstrate cooling throughout the Mesozoic (Fig. 13). However, this is related to sensitivity of the AFT method. The samples (B38, B45) that were at depths within the apatite partial annealing zone at the beginning of Mesozoic were unable to record early Mesozoic re-heating.

Geological implications
A deep erosional incision within the sedimentary cover of the East European Platform has been conventionally attributed to the effects of the Variscan orogeny (e.g., Żelichowski, 1987;Narkiewicz, 2007). This approach was consistent with wide-scale observations of a regional early Stephanian hiatus over much of the Variscan foreland with Stephanian and lower Permian red beds unconformably overlying truncated Westphalian series (e.g., McCann, 1996). However, our models indicate early Car-570 boniferous exhumation of the lower Palaeozoic strata in the Baltic and Podlasie Basins of NE Poland on the SW slope of the EEP. This agrees with the presence of Carboniferous sediments resting on the deeply eroded Silurian substratum over a sig-and lasted until latest Carboniferous Variscan shortening ( ( Fig. 2c; Krzywiec et al., 2017a). Recent seismic data indicate a Variscan thin-skinned fold-and-thrust belt emplaced on SW slope of the EEP within the Lublin Basin and west of it (Krzywiec et al., 2017a, 20017b). Hence, latest Carboniferous heating might have been caused by a combined effect of tectonic and sedimentary burial ( Fig. 5b; Botor et al., 2019a). A significant cover of the Variscan foreland basin was also inferred byŚrodoń et al. (2013) for western Ukraine based on combined clay mineralogy, K-Ar, and AFT data.

615
From the beginning of Permian, the SW slope of the EEP was onlapped by marginal part of an extensive Permian-Mesozoic basin (Figs. 2, 4). Considering a present-day Permian-Mesozoic thickness (Figs. 2, 4), i.e., omitting effects of the Late Cretaceous inversion, a thermal effect of Mesozoic burial should be significant. Nevertheless, among the methods used, only the AFT thermochronometer is sensitive enough to record this event. Furthermore, only models for Permian-Mesozoic samples are capable to reveal a Mesozoic thermal history. Consequently, only three models predict Mesozoic re-heating with a max-620 imum paleotemperature at the transition from Triassic to Jurassic (Fig. 15), but this result is probably representative for the entire study area. Comparable results were obtained by Schito et al. (2018) in the Ukrainian part of the EEP, who postulated that exhumation through the 45-120°C temperature range took place between the Late Triassic and Early Jurassic, and that no significant burial occurred afterwards.

625
The case of the SW slope of the EEP demonstrates a successful application of integrative approach to studying a long-term thermal history. A series of tectonic events in the area has not been yet fully resolved because of incomplete sedimentary record. Therefore, the combination of ZHe and AFT termochronometers with VR data was necessary to recognise those events that led to deep erosion and complete removal of sedimentary sequences. The same approach might be applied to craton margins elsewhere to separate superimposed effects of successive tectonic events.

630
Significant heating of Ediacaran to Carboniferous sedimentary successions occurred before the Permian with maximum paleotemperatures in the earliest and latest Carboniferous in the Baltic-Podlasie Basin and Lublin Basin, respectively. The results obtained suggest an important role of early Carboniferous uplift and exhumation at the SW margin of the EEP. This event was associated with a period of intra-plate magmatism in the area (Poprawa, 2019) lasting from the late Tournaisian to mid-Viséan (Pańczyk and Nawrocki, 2015). Effects of uplift and magmatism jointly suggest thermal perturbation of lithosphere.

635
The time of this event was roughly coeval with the extensional reactivation of the Dniepr-Donets-Donbas Rift (e.g., Stephenson et al., 2006) and termination of the Late Devonian continental rifting in the Pripyat Trough (Kusznir et al., 1996).
The SW slope of the EEP was overridden in SE Poland (Lublin Basin) by the Variscan orogenic wedge (Krzywiec et al., 2017a;Mazur et al., 2020). This event interrupted Carboniferous uplift because of tectonic loading and caused resumption of on the y-scale (illustrated as ±2σ). The age of each crystal can be determined by a line from the origin through the projection point of the crystal to the intercept on the radial age scale (Galbraith, 1990)