Neanderthal coexistence with Homo sapiens in Europe was affected by herbivore carrying capacity

Abstract

It has been proposed that climate change and the arrival of modern humans in Europe affected the disappearance of Neanderthals due to their impact on trophic resources; however, it has remained challenging to quantify the effect of these factors. By using Bayesian age models to derive the chronology of the European Middle to Upper Paleolithic transition, followed by a dynamic vegetation model that provides the Net Primary Productivity, and a macroecological model to compute herbivore abundance, we show that in continental regions where the ecosystem productivity was low or unstable, Neanderthals disappeared before or just after the arrival of Homo sapiens. In contrast, regions with high and stable productivity witnessed a prolonged coexistence between both species. The temporal overlap between Neanderthals and H. sapiens is significantly correlated with the carrying capacity of small- and medium-sized herbivores. These results suggest that herbivore abundance released the trophic pressure of the secondary consumers guild, which affected the coexistence likelihood between both human species.

Fig. 1.. Net primary productivity.

(A) Dendrogram clustering plot according to the NPP differences obtained from the dCORT dissimilarity index. Each color indicates one cluster. (B) Archaeo-paleontological sites where NPP was estimated. Colors correspond with the same colors used in the dendrogram plot. (C) Temporal evolution of NPP in each archaeo-paleontological site grouped according to the clusters obtained from the dCORT index. The colored line represents the mean, corresponding with the color used in (A) and (B).

Fig. 2.. Climate and herbivore species during MIS3.

(A) Köppen-Geiger climate classification in each archaeo-paleontological site included in this study. D, continental; C, temperate; E, polar; T, tundra; w, dry winter; f, no dry season; s, dry summer; a, hot summer; b, warm summer; c, cold summer. All climate variables were obtained from the work of Armstrong et al. (90) after bias correction (see the “Paleoclimate reconstruction and validation” section for details). (B) Cluster dendrogram with the classification of each herbivore paleocommunity in each region according to the Jaccard Dissimilarity Index (JDI). The figure depicts the herbivore species recovered in each region during MIS3.

Fig. 3.. Biogeographic regions of Europe.

Biogeographic regions during MIS3 according to the climate conditions, the NPP, and the herbivore guild composition. (A) Dots represent the archaeo-paleontological sites included in the study. (B) Mean CC (kg/km² per year) of large-sized (>300 kg), medium-sized (20 to 300 kg), and small-sized (<20 kg) herbivores in each region during MIS3. (C) Temporal evolution of the NPP during MIS3 in each biogeographic region. Vertical numbered bars represent the stadial phases. Horizontal bars show the timing of the NAI (in purple) disappearance at 68% CI (with bold, thicker bar) and 95% CI, and the timing of the SAI (in green) appearance at 68% and 95% CI. Asterisks (*) indicate that the BAMs were run with the most restrictive filtering criteria (filtering criteria 2; for details, see the “Chronological models” section in Materials and Methods).

Fig. 4.. Ecosystem productivity and coexistence between Neanderthals and H. sapiens.

Box and whisker plots showing the NPP, its coefficient of variation (CV), and the CC of small-, medium-, and large-sized herbivores in those regions with (Coexistence) and without (No coexistence) temporal overlap between Neanderthals and H. sapiens. The center line of each plot corresponds to the median. The box edges correspond to the first and third quartiles. The whiskers correspond to 1.5× the interquartile range, and the dots represent outliers.

Fig. 5.. Correlation between dependent and independent variables.

Dependent variables are as follows: the timing of the NAI disappearance, the SAI appearance, and the temporal overlap between the NAI and the SAI in each region. Independent variables are NPP and CC of small-, medium-, and large-sized herbivores. Each ESF test was performed with the chronometric filtering criteria that produced the FC 1 and 2 (for details, see “Chronological models” in Materials and Methods). Violin and box plots show the range of P values after the resampling procedure that incorporates the uncertainty at 95% CI of each dependent and independent variable. The vertical dashed red line indicates the threshold of 0.05 in the P value, so values to the left (P ≤ 0.05) are statistically significant.

Fig. 6.. Comparison between OLE and BAM outputs.

(A) Violin plot with the mean and the 95% CI for the end of the NAI (blue color) and the beginning of the SAI (red color) in each region according to OLEs. (B) Mean and 95% CI interval for the end of the NAI (in blue color) and the start of the SAI (red color) according to the BAMs. (C) Linear correlation between the chronologies obtained with BAM and OLE models. Each point represents one biogeographic region identified with the label code.

Fig. 7.. Sensitivity to the culture-species association uncertainty.

Correlation between dependent and independent variables when (A) the Châtelperronian is excluded from the NAI and the Uluzzian is excluded from the SAI and (B) the Uluzzian is incorporated into the NAI. Each ESF test was performed with the chronometric FC 1 and 2. Violin and box plots show the range of P values after the resampling procedure that incorporates the uncertainty at 95% CI of each dependent and independent variable. The vertical dashed red line indicates the threshold of 0.05 in the P value, so values to the left (P < 0.05) are statistically significant.