Article Search
Volume 27.1
January–April 2024
Full table of contents
ISSN: 1094-8074, web version;
1935-3952, print version
Recent Research Articles
See all articles in 27.1 January-April 2024
See all articles in 26.3 September-December 2023
See all articles in 26.2 May-August 2023
See all articles in 26.1 January-April 2023
Esther Galbrun. Department of Computer Science, Aalto University, P.O. Box 15400, FI-00076 Aalto, Finland. esther.galbrun@aalto.fi
Esther Galbrun works mostly on data mining methods, focusing in particular on redescription mining. She is a postdoctoral researcher at the Department of Computer Science, Aalto University, Finland.
Hui Tang. Department of Geosciences, University of Oslo, P.O. Box 1022, University of Oslo, 0315-Oslo, Norway. hui.tang@geo.uio.no
Hui Tang is a Postdoc in Department of Geoscience, University of Oslo, Norway. He works with global and regional climate models to better understand climate changes in the past and future. He is also developing dynamic vegetation model for the Arctic region to better depict vegetation changes and their climatic feedbacks in the region.
Mikael Fortelius. Department of Geosciences and Geography, University of Helsinki, P.O. Box 64, FI-00014 University of Helsinki, Finland. mikael.fortelius@helsinki.fi
Mikael Fortelius is a palaeontologist with special interest in the relationships between climate, vegetation and herbivores. He has a long-standing interest in how mammalian teeth work, grow, and evolve. He is Professor of Evolutionary Palaeontology in the Department of Geoscience and Geography at the University of Helsinki. For the last 20 years he has been engaged in developing and coordinating the NOW database of fossil mammals (http://www.helsinki.fi/science/now/).
Indrė Žliobaitė. Department of Computer Science, University of Helsinki, P.O. Box 68, FI-00014 University of Helsinki, Finland; Department of Geosciences and Geography, University of Helsinki, P.O. Box 64, FI-00014 University of Helsinki, Finland. indre.zliobaite@helsinki.fi
Indre Zliobaite is an Assistant Professor at the Department of Computer Science, University of Helsinki, Finland, where she leads a research group on data science and evolution (http://www.helsinki.fi/data-science-and-evolution).
TABLE 1. Dental trait variables organized by categories.
Teeth durability | |||
Hypsodonty (HYP): | brachydont (1) | mesodont (2) | hypsodont (3) |
Horizodonty (HOD): | brachyhorizodont (1) | mesohorizodont (2) | hypsohorizodont (3) |
Cutting structures | |||
Acute lophs (AL): | absent (0) | present (1) | |
Obtuse or basin-like lophs (OL): | absent (0) | present (1) | |
Occlusion characteristics | |||
Structural fortification of cusps (SF): | absent (0) | present (1) | |
Occlusal topography (OT): | has raised elements (0) | is flat (1) | |
Material properties | |||
Coronal cementum (CM): | absent or very thin (0) | thick coating (1) |
TABLE 2. Number of sites from each continent containing taxa from the given order or family, after (left) and before (right) filtering out sites with fewer than three taxa.
Eurasia | Africa | North America | South America | |||||
Total no. of sites | 12497 | 21586 | 8235 | 12029 | 2544 | 9636 | 5610 | 7113 |
Artiodactyla | 12467 | 20269 | 8131 | 11637 | 2544 | 8599 | 5610 | 6292 |
Antilocapridae | 0 | 0 | 0 | 0 | 699 | 809 | 0 | 0 |
Bovidae | 6278 | 9037 | 8125 | 11545 | 726 | 1465 | 0 | 0 |
Camelidae | 0 | 0 | 0 | 0 | 0 | 0 | 52 | 301 |
Cervidae | 10903 | 15566 | 45 | 76 | 2544 | 8358 | 5582 | 6138 |
Giraffidae | 0 | 0 | 957 | 957 | 0 | 0 | 0 | 0 |
Hippopotamidae | 0 | 0 | 772 | 772 | 0 | 0 | 0 | 0 |
Moschidae | 3980 | 4129 | 0 | 0 | 0 | 0 | 0 | 0 |
Suidae | 8692 | 11215 | 6222 | 6356 | 0 | 0 | 0 | 0 |
Tayassuidae | 0 | 0 | 0 | 0 | 545 | 862 | 5527 | 5642 |
Tragulidae | 983 | 985 | 997 | 997 | 0 | 0 | 0 | 0 |
Perissodactyla | 876 | 888 | 2855 | 2855 | 297 | 297 | 5277 | 5293 |
Equidae | 837 | 849 | 990 | 990 | 0 | 0 | 0 | 0 |
Rhinocerotidae | 5 | 5 | 2800 | 2800 | 0 | 0 | 0 | 0 |
Tapiridae | 35 | 35 | 0 | 0 | 297 | 297 | 5277 | 5293 |
Primates | 4146 | 4555 | 7704 | 8002 | 356 | 356 | 5320 | 5349 |
Aotidae | 0 | 0 | 0 | 0 | 0 | 0 | 622 | 622 |
Atelidae | 0 | 0 | 0 | 0 | 350 | 350 | 273 | 273 |
Callitrichidae | 0 | 0 | 0 | 0 | 0 | 0 | 2638 | 2638 |
Cebidae | 0 | 0 | 0 | 0 | 243 | 243 | 5311 | 5340 |
Cercopithecidae | 4146 | 4552 | 7582 | 7841 | 0 | 0 | 0 | 0 |
Cheirogaleidae | 0 | 0 | 88 | 109 | 0 | 0 | 0 | 0 |
Daubentoniidae | 0 | 0 | 49 | 51 | 0 | 0 | 0 | 0 |
Galagidae | 0 | 0 | 5810 | 5823 | 0 | 0 | 0 | 0 |
Hominidae | 53 | 53 | 1193 | 1193 | 0 | 0 | 0 | 0 |
Hylobatidae | 841 | 841 | 0 | 0 | 0 | 0 | 0 | 0 |
Indridae | 0 | 0 | 64 | 70 | 0 | 0 | 0 | 0 |
Lemuridae | 0 | 0 | 77 | 95 | 0 | 0 | 0 | 0 |
Lepilemuridae | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
Lorisidae | 1186 | 1196 | 1585 | 1585 | 0 | 0 | 0 | 0 |
Megaladapidae | 0 | 0 | 32 | 35 | 0 | 0 | 0 | 0 |
Pitheciidae | 0 | 0 | 0 | 0 | 0 | 0 | 1548 | 1548 |
Tarsiidae | 393 | 407 | 0 | 0 | 0 | 0 | 0 | 0 |
Proboscidea | 245 | 245 | 2559 | 2559 | 0 | 0 | 0 | 0 |
Elephantidae | 245 | 245 | 2559 | 2559 | 0 | 0 | 0 | 0 |
TABLE 3. List of the dental traits and bioclimatic variables. Temperature and precipitation are measured respectively in degrees Celsius (°C) and in millimeters (mm).
Dental trait variables | |
Hyp1 | Fraction of brachydont taxa (Hypsodonty) |
Hyp2 | Fraction of mesodont taxa (Hypsodonty) |
Hyp3 | Fraction of hypsodont taxa (Hypsodonty) |
Hod1 | Fraction of brachyhorizodont taxa (Horizodonty) |
Hod2 | Fraction of mesohorizodont taxa (Horizodonty) |
Hod3 | Fraction of hypsohorizodont taxa (Horizodonty) |
AL | Fraction of taxa with acute lophs |
OL | Fraction of taxa with obtuse lophs |
SF | Fraction of taxa with structural fortification of cups |
OT | Fraction of taxa with flat occlusal topography |
CM | Fraction of taxa with coronal cementum |
Bioclimatic variables | |
T~Y | Mean Annual Temperature |
T~RngD | Mean Diurnal Range |
TIso | Isothermality |
TSeason | Temperature Seasonality |
T+WarmM | Max Temperature of Warmest Month |
T-ColdM | Min Temperature of Coldest Month |
TRngY | Annual Temperature Range |
T~WetQ | Mean Temperature of Wettest Quarter |
T~DryQ | Mean Temperature of Driest Quarter |
T~WarmQ | Mean Temperature of Warmest Quarter |
T~ColdQ | Mean Temperature of Coldest Quarter |
PTotY | Annual Precipitation |
PWetM | Precipitation of Wettest Month |
PDryM | Precipitation of Driest Month |
PSeason | Precipitation Seasonality |
PWetQ | Precipitation of Wettest Quarter |
PDryQ | Precipitation of Driest Quarter |
PWarmQ | Precipitation of Warmest Quarter |
PColdQ | Precipitation of Coldest Quarter |
TABLE 4. Extending a redescription: an example in four steps.
qD | qC | J | |supp| |
[1≤ OL ≤1] | [T-ColdM ≤-10.3] | 0.55 | 4825 |
[1≤ OL ≤1] OR [0.4≤ SF ≤0.4] | [T-ColdM ≤-10.3] | 0.6 | 5615 |
([1≤ OL ≤1] OR [0.4≤ SF ≤0.4]) AND [AL ≤0] | [T-ColdM ≤-10.3] | 0.61 | 5615 |
([1≤ OL ≤1] OR [0.4≤ SF ≤0.4]) AND [AL ≤0] | [T-ColdM ≤-10.3] AND [0.1≤ T~WarmQ ≤21.6] | 0.62 | 5472 |
TABLE 5. Ten redescriptions with highest accuracy among 379 obtained. For each redescription, we list its queries, that is, the query over dental traits variables (qD) and the query over bioclimatic variables (qC). We also indicate the accuracy of the redescription (J) as well as the size of its support, as the number of sites described (|supp|) and as a percentage of the total number of sites (supp%).
R1 | J = 0.68 | |supp| = 6517 | supp% = 22.56 | ||
qD = [Hyp2 ≤0.333] AND [1≤ Hod1] AND [AL ≤0.056] AND [0.75≤ OL] | |||||
qC = [T~WarmQ ≤18.3] AND [T~ColdQ ≤6] | |||||
R2 | J = 0.67 | |supp| = 6694 | supp% = 23.17 | ||
qD = ( ( [0.846≤ Hyp1] AND [OL ≤0.4] ) OR [0.033≤ OT ≤0.138] ) AND [Hyp3 ≤0.348] | |||||
qC = [67≤ TIso] AND [17.7≤ T+WarmM ≤35.8] | |||||
R3 | J = 0.65 | |supp| = 6291 | supp% = 21.78 | ||
qD = [Hyp2 ≤0.333] AND [1≤ Hod1] AND [AL ≤0.048] AND [0.75≤ OL] | |||||
qC = [T+WarmM ≤25.7] AND [T~ColdQ ≤6.1] | |||||
R4 | J = 0.63 | |supp| = 5470 | supp% = 18.94 | ||
qD =( [1≤ OL ≤1] OR [0.4≤ SF ≤0.4] ) AND [Hyp2 ≤0.333] AND [AL ≤0] | |||||
qC =[T-ColdM ≤-10.3] AND [0.1≤ T~WarmQ ≤21.6] | |||||
R5 | J = 0.63 | |supp| = 6374 | supp% = 22.07 | ||
qD =( ( [Hyp3 ≤0.458] AND [0.061≤ AL ≤0.235] ) OR [0.032≤ Hod3 ≤0.059] ) AND [OL ≤0.643] | |||||
qC =[68≤ TIso ≤91] AND [613≤ PTotY ≤6989] | |||||
R6 | J = 0.62 | |supp| = 4821 | supp% = 16.69 | ||
qD =( [1≤ Hyp1 ≤1] AND [0.091≤ SF ≤0.286] ) OR [0.033≤ Hyp3 ≤0.12] OR [0.182≤ AL ≤0.188] | |||||
qC =[53≤ TIso ≤91] AND [1475≤ PTotY ≤3670] | |||||
R7 | J = 0.61 | |supp| = 3476 | supp% = 12.03 | ||
qD =( [1≤ Hyp1 ≤1] AND [0.059≤ OL ≤0.333] AND [0.091≤ SF ≤0.25] ) OR [0.062≤ Hyp2 ≤0.083] | |||||
qC =[65≤ TIso ≤91] AND [692≤ PWetQ ≤1511] | |||||
R8 | J = 0.60 | |supp| = 4971 | supp% = 17.21 | ||
qD =( ( [1≤ Hyp1 ≤1] AND [OL ≤0.333] ) OR [0.033≤ OT ≤0.107] ) AND [0.032≤ AL ≤0.188] | |||||
qC =[23.1≤ TSeason ≤116.6] AND [289≤ PWetQ ≤2256] | |||||
R9 | J = 0.60 | |supp| = 4666 | supp% = 16.15 | ||
qD =( ( [0.933≤ Hyp1] AND [0.059≤ OL ≤0.364] ) OR [0.033≤ OT ≤0.107] ) AND [0.032≤ AL ≤0.188] | |||||
qC =[69≤ TIso ≤87] AND [410≤ PWetQ ≤1940] | |||||
R10 | J = 0.60 | |supp| = 5993 | supp% = 20.75 | ||
qD =( ( [0.759≤ OL] AND [CM ≤0] ) OR [0.5≤ SF ≤0.667] ) AND [0.25≤ Hyp1] | |||||
qC =[TIso ≤31] AND [T~ColdQ ≤2.2] |
TABLE 6. Redescriptions R1 and variants with alternative dental traits queries. R1a is obtained by manually removing Hyp2. The remaining variants are obtained by deleting the entire query, then letting the algorithm find a new one, with some variables deactivated. First, we deactivated variables Hod1, Hod2, Hod3 and AL, obtaining R1b. Further deactivating Hyp2, we obtained R1c and R1d. Fields are the same as in Table 5.
R1 | J = 0.68 | |supp| = 6517 | supp% = 22.56 | ||
qD =[Hyp2 ≤0.333] AND [1≤ Hod1] AND [AL ≤0.056] AND [0.75≤ OL] | |||||
qC =[T~WarmQ ≤18.3] AND [T~ColdQ ≤6] | |||||
R1a | J = 0.67 | |supp| = 6518 | supp% = 22.56 | ||
qD =[1≤ Hod1] AND [AL ≤0.056] AND [0.75≤ OL] | |||||
qC =[T~WarmQ ≤18.3] AND [T~ColdQ ≤6] | |||||
R1b | J = 0.61 | |supp| = 6532 | supp% = 22.61 | ||
qD =[Hyp2 ≤0.333] AND [0.75≤ OL] | |||||
qC =[T~WarmQ ≤18.3] AND [T~ColdQ ≤6] | |||||
R1c | J = 0.65 | |supp| = 6286 | supp% = 21.76 | ||
qD =( [0.75≤ OL] AND [CM ≤0] ) OR [0.8≤ Hyp3] | |||||
qC =[T~WarmQ ≤18.3] AND [T~ColdQ ≤6] | |||||
R1d | J = 0.66 | |supp| = 6158 | supp% = 21.32 | ||
qD =( [0.273≤ Hyp1] AND [0.75≤ OL] AND [CM ≤0] ) OR [0.8≤ Hyp3] | |||||
qC =[T~WarmQ ≤18.3] AND [T~ColdQ ≤6] |
TABLE 7. Redescriptions R2 and R5 and variants with alternative dental traits queries. R2a and R2b are obtained by splitting the dental traits queries of R2 into two components. Similarly, R5a and R5b are obtained by splitting dental traits queries of R5 into two components. Fields are the same as in Table 5.
R2 | J = 0.67 | |supp| = 6694 | supp% = 23.17 | ||
qD =( ( [0.846≤ Hyp1] AND [OL ≤0.4] ) OR [0.033≤ OT ≤0.138] ) AND [Hyp3 ≤0.348] | |||||
qC =[67≤ TIso] AND [17.7≤ T+WarmM ≤35.8] | |||||
R2a | J = 0.52 | |supp| = 4979 | supp% = 17.24 | ||
qD =[0.846≤ Hyp1] AND [OL ≤0.4] | |||||
qC =[67≤ TIso] AND [17.7≤ T+WarmM ≤35.8] | |||||
R2b | J = 0.20 | |supp| = 1758 | supp% = 6.09 | ||
qD =[0.033≤ OT ≤0.138] AND [Hyp3 ≤0.348] | |||||
qC =[67≤ TIso] AND [17.7≤ T+WarmM ≤35.8] | |||||
R5 | J = 0.63 | |supp| = 6374 | supp% = 22.07 | ||
qD =( ( [Hyp3 ≤0.458] AND [0.061≤ AL ≤0.235] ) OR [0.032≤ Hod3 ≤0.059] ) AND [OL ≤0.643] | |||||
qC =[68≤ TIso ≤91] AND [613≤ PTotY ≤6989] | |||||
R5a | J = 0.57 | |supp| = 5604 | supp% = 19.40 | ||
qD =[0.061≤ AL ≤0.235] AND [OL ≤0.643] | |||||
q C =[68≤ TIso ≤91] AND [613≤ PTotY ≤6989] | |||||
R5b | J = 0.14 | |supp| = 1077 | supp% = 3.73 | ||
qD =[0.032≤ Hod3 ≤0.059] AND [OL ≤0.643] | |||||
qC =[68≤ TIso ≤91] AND [613≤ PTotY ≤6989] |
TABLE 8. Three redescriptions, R43, R69 and R74, featuring high values of PSeason with highest accuracy among 379 obtained and variants with alternative dental traits queries. R43a and R43b are obtained from R43 by removing first Hod1, then OL. R69a and R69b are obtained by splitting the dental traits queries of R69 into two components. Fields are the same as in Table 5.
R43 | J = 0.52 | |supp| = 3141 | supp% = 10.87 | ||
qD =( [Hyp1 ≤0.429] AND [0.042≤ Hod3 ≤0.222] ) OR [0.96≤ Hod1 ≤0.96] OR [0.167≤ OL ≤0.167] | |||||
qC =[54≤ TIso ≤88] AND [84≤ PSeason ≤136] | |||||
R43a | J = 0.51 | |supp| = 3101 | supp% = 10.74 | ||
qD =( [Hyp1 ≤0.429] AND [0.042≤ Hod3 ≤0.222] ) OR [0.167≤ OL ≤0.167] | |||||
qC =[54≤ TIso ≤88] AND [84≤ PSeason ≤136] | |||||
R43b | J = 0.50 | |supp| = 2864 | supp% = 9.91 | ||
qD =[Hyp1 ≤0.429] AND [0.042≤ Hod3 ≤0.222] | |||||
qC =[54≤ TIso ≤88] AND [84≤ PSeason ≤136] | |||||
R69 | J = 0.48 | |supp| = 5073 | supp% = 17.56 | ||
qD =( [0.346≤ Hyp3] AND [AL ≤0.095] ) OR [0.444≤ Hyp2] OR [0.429≤ SF ≤0.455] | |||||
qC =[13.8≤ TRngY ≤50.3] AND [91≤ PSeason ≤164] | |||||
R69a | J = 0.45 | |supp| = 4640 | supp% = 16.06 | ||
qD =[0.346≤ Hyp3] AND [AL ≤0.095] | |||||
qC =[13.8≤ TRngY ≤50.3] AND [91≤ PSeason ≤164] | |||||
R69b | J = 0.03 | |supp| = 197 | supp% = 0.68 | ||
qD =[0.429≤SF≤0.455] | |||||
qC =[13.8≤ TRngY ≤50.3] AND [91≤ PSeason ≤164] | |||||
R74 | J = 0.47 | |supp| = 5070 | supp% = 17.55 | ||
qD =( [Hyp1 ≤0.4] OR [0.111≤ Hyp2 ≤0.125] OR [0.273≤ Hyp3 ≤0.308] ) AND [AL ≤0.095] | |||||
qC =[75.1≤ TSeason ≤1352] AND [90≤ PSeason ≤147] |
TABLE 9. Climate classes defined by the Köppen system (Kottek et al., 2006). Pth is a dryness threshold.
Class | Description | Climate criterion |
A | Equatorial climates | Tmin ≤18°C |
Af | Equatorial rainforest, fully humid | Pmin ≥60 mm |
Am | Equatorial monsoon | Pann ≥25(100 - P min ) |
As | Equatorial savanna with dry summer | Pmin ≤60 mm in summer |
Aw | Equatorial savanna with dry winter | Pmin ≤60 mm in winter |
B | Arid climates | Pann <10 Pth |
BS | Steppe climate | Pann >5 Pth |
BW | Desert climate | Pann ≤5 Pth |
C | Warm temperate climates | -3°C < T min < +18°C |
Cs | Warm temperate climate with dry summer | Psmin < Pwmin, Pwmax > 3 Psmin and Psmin < 40 mm |
Cw | Warm temperate climate with dry winter | Pwmin < Psmin and Psmax > 10 P wmin |
Cf | Warm temperate climate, fully humid | neither Cs nor Cw |
D | Snow climates | Tmin ≤-3°C |
Ds | Snow climate with dry summer | Psmin < Pwmin, Pwmax > 3 Psmin and Psmin < 40 mm |
Dw | Snow climate with dry winter | Pwmin < Psmin and smax > 10 Pwmin |
Df | Snow climate, fully humid | neither Ds nor Dw |
E | Polar climates | Tmax < + 10°C |
ET | Tundra climate | 0°C ≤Tmax < +10°C |
EF | Frost climate | Tmax < 0°C |
FIGURE 1. Examples of functional crown type scores, each row presents a set of teeth with different occlusal topography. Tooth sizes are not to scale. 1, Diceros ; 2, Listriodon ; 3, Giraffa ; 4, Pan ; 5, Megaladapis ; 6, Kobus ; 7, Hippotragus ; 8, Hippopotamus ; 9, Hylochoerus ; 10, Ceratotherium ; 11, Loxodonta ; 12, Bos ; 13, Equus ; and 14, Phacochoerus . The figure has been adapted from Zliobaite et al. (2016). Sources of the illustrations: Diceros and Ceratotherium are from figure 2 in Fortelius (1981), Kobus and Hyppotragus are from figure 2 in Kaiser et al. (2010), all the other examples come from several illustrations in Thenius (1989).
FIGURE 2. Datasets, data aggregation and mining processes. The initial datasets (Sites × Taxa) and (Taxa × Traits) are aggregated to produce the (Sites × Traits) dataset. Redescriptions are then mined from this dataset and the (Sites × Climate) dataset, resulting in a collection of redescriptions denoted as R1, R2, etc.
FIGURE 3. Maps of global dental trait distributions. Each site is represented as a colored square on the map. Next to each plot, a colorbar indicates the mapping from colors to traits values (right side of the legend) and a histogram depicts the distribution of those values (left side of the legend).
FIGURE 4. Maps of bioclimatic variables: temperatures. Each site is represented as a colored square on the map. Next to each plot, a colorbar indicates the mapping from colors to the values of the temperature variables (right side of the legend) and a histogram depicts the distribution of those values (left side of the legend).
FIGURE 5. Maps of bioclimatic variables: precipitation. Each site is represented as a colored square on the map. Next to each plot, a colorbar indicates the mapping from colors to the values of the precipitation variables (right side of the legend) and a histogram depicts the distribution of those values (left side of the legend).
FIGURE 6. Maps of redescriptions R1 to R9. Locations that satisfy both queries of the redescription are plotted in dark purple (darkest shade of gray), locations that satisfy only the dental traits query and only the climate query are plotted in red and blue, respectively (intermediate shades of gray), while locations that satisfy neither queries are plotted in light gray.
FIGURE 7. Maps of redescriptions R1 and its variants. Locations that satisfy both queries of the redescription are plotted in dark purple (darkest shade of gray), locations that satisfy only the dental traits query and only the climate query are plotted in red and blue, respectively (intermediate shades of gray), while locations that satisfy neither queries are plotted in light gray.
FIGURE 8. Maps of redescriptions R2, R5 and variants. Locations that satisfy both queries of the redescription are plotted in dark purple (darkest shade of gray), locations that satisfy only the dental traits query and only the climate query are plotted in red and blue, respectively (intermediate shades of gray), while locations that satisfy neither queries are plotted in light gray.
FIGURE 9. Maps of redescriptions R43, R69, R74 and variants. Locations that satisfy both queries of the redescription are plotted in dark purple (darkest shade of gray), locations that satisfy only the dental traits query and only the climate query are plotted in red and blue, respectively (intermediate shades of gray), while locations that satisfy neither queries are plotted in light gray.
FIGURE 10. Figures comparing the supports of the redescriptions to the Köppen-Geiger climate classification system. 1, Map of the distribution the Köppen climate subclasses in our dataset. 2, Histograms showing the distribution of the support of the redescriptions over those subclasses and entropy ratio evaluating the match between the support and the subclasses.
Computational biomes: The ecometrics of large mammal teeth
Plain Language Abstract
Plant eating animals obtain their energy from plant food, which they process using their teeth. Different kinds of plants grow in different environments. Therefore, species surviving in different environments require different kinds of teeth. Here we use data mining techniques to analyze what types of teeth associate with what environments.
Resumen en Español
Biomas computacionales: la ecometría de los dientes de grandes mamíferos
Como los organismos están adaptados a sus entornos, las asociaciones de taxones se pueden utilizar para describir entornos del presente y del pasado. Aquí se utiliza un método de extracción de datos (data mining), concretamente extracción de redescripción (redescription mining), para descubrir y analizar patrones de asociación entre grandes mamíferos herbívoros y sus entornos a partir de sus características funcionales. Nos centramos en las propiedades funcionales de los dientes de animales, caracterizadas dentro de un esquema de puntuación del rasgo dental que ha sido recientemente desarrollado. Los dientes de los mamíferos herbívoros sirven de medio para obtener energía de los alimentos y, por lo tanto, se espera que coincidan con los tipos de alimentos vegetales disponibles en sus entornos. Según esto, se espera que los rasgos dentales proporcionen evidencias de las condiciones ambientales. Analizamos un conjunto de datos global de los registros de grandes mamíferos herbívoros y de condiciones bioclimáticas. Identificamos patrones comunes de asociación entre las distribuciones de los rasgos dentales y las condiciones bioclimáticas y discutimos sus implicaciones. Cada patrón se puede considerar como un bioma computacional. Nuestro análisis distingue tres zonas globales, a las que nos referimos como la zona húmeda boreal-templada, la zona húmeda tropical y la zona seca tropical-subtropical. La zona húmeda boreal-templada se caracteriza principalmente por temperaturas frías estacionales, una falta de hipsodoncia y una alta proporción de especies con lofos obtusos. La zona húmeda tropical se caracteriza principalmente por altas temperaturas, alta isotermia, abundante precipitación y una gran proporción de especies con lofos agudos en lugar de obtusos. Finalmente, la zona seca tropical se caracteriza principalmente por una alta estacionalidad de las temperaturas y las precipitaciones, así como una alta hipsodoncia y horizodoncia (horizodonty). Encontramos que la marca dental para los bosques tropicales africanos es bastante diferente de la marca dental de sitios climáticamente similares en América del Norte y Asia, donde las especies hipsodontas y las especies con lofos obtusos están ausentes en su mayoría. En términos climáticos y de marcas dentales, los trópicos estacionales africanos comparten muchas similitudes con las localidades de Asia Central y del Sur. Curiosamente, la meseta tibetana está cubierta tanto por redescripciones del grupo seco tropical-subtropical como por redescripciones del grupo húmedo boreal-templado, lo que sugiere una combinación de características de ambas zonas en sus características dentales y en el clima.
Palabras clave: ecometría; extracción de redescripción; rasgos dentales; grandes mamíferos; extracción de datos
Traducción: Enrique Peñalver (Sociedad Española de Paleontología)
Résumé en Français
Biomes statistiques : les traits indicateurs d’écologie des dents de grands mammifères
Comme les organismes sont adaptés à leurs environnements, des assemblages de taxons peuvent être utilisés pour décrire les environnements dans le présent et dans le passé. Dans cet article, nous utilisons une méthode d’exploration des données, la fouille de redescriptions, pour découvrir et analyser des schémas d’association des grands mammifères herbivores et de leurs environnements via leurs traits fonctionnels. Nous nous concentrons sur les propriétés fonctionnelles des dents des animaux, caractérisées en utilisant un système de classification des traits dentaires récemment développé. Les dents des mammifères herbivores servent d’interface pour obtenir l’énergie à partir de la nourriture, et il est donc attendu qu’elles correspondent aux types de plantes disponibles dans leur environnement. Ainsi, il est attendu que les traits dentaires portent un signal des conditions environnementales. Nous analysons une compilation globale des occurrences de grands mammifères herbivores et de leurs conditions bioclimatiques. Nous identifions des schémas communs d’association entre les distributions des traits dentaires et les conditions bioclimatiques et nous discutons leurs implications. Chaque schéma peut être considéré comme un biome statistique. Notre analyse distingue trois zones globales : la zone humide boréale-tempérée, la zone humide tropicale, et la zone sèche tropicale-subtropicale. La zone humide boréale-tempérée est principalement caractérisée par des températures froides et saisonnières, un manque d’hypsodontie, et une forte proportion d’espèces avec des lophes obtus. La zone humide tropicale est principalement caractérisée par des températures élevées, une isothermalité forte, des précipitations abondantes, et une forte proportion d’espèces avec des lophes aigus plutôt qu’obtus. Enfin, la zone sèche tropicale est principalement caractérisée par une saisonnalité forte des températures et des précipitations, ainsi qu’une hypsodontie et une horizodontie fortes. Nous observons que la signature des traits dentaires des forêts tropicales humides africaines est assez différente de celle des sites d’Amérique du Nord et d’Asie situés sous des climats similaires, sites dans lesquels les espèces hypsodontes et les espèces avec des lophes obtus sont généralement absentes. En termes de signatures dentaires et climatiques, les tropiques saisonniers en Afrique partagent de nombreuses similarités avec les sites d’Asie centrale et d’Asie du Sud. De manière intéressante, le plateau tibétain est couvert à la fois de redescriptions du groupe tropical-subtropical sec et de redescriptions du groupe boréal-tempéré humide, suggérant une combinaison de caractères des deux zones dans les traits dentaires et le climat.
Mots-clés : traits indicateurs d’écologie ; fouille de redescriptions ; traits dentaires ; grands mammifères ; exploration des données
Translator: Antoine Souron
Deutsche Zusammenfassung
Rechnergestützte Biome: die Ökometrie großer Säugetierzähne
Da Organismen ihrer Umwelt angepasst sind, können vergangene und heutige Milieus anhand von Taxa-Assemblagen beschrieben werden. In dieser Untersuchung arbeiten wir mit einem Data-Mining-Verfahren, dem Redescription Mining, mit dem wir Assoziationsmuster zwischen großen herbivoren Säugern und deren Umwelt über funktionale Merkmale herausfinden und analysieren. Der Fokus liegt auf den funktionellen Eigenschaften von Tierzähnen, die durch ein kürzlich entwickeltes Merkmalbewertungsschema charakterisiert wurden. Die Zähne herbivorer Säugetiere sind die Schnittstelle zur Energiegewinnung aus Nahrung und daher wird erwartet, dass sie mit den in der Umgebung vorkommenden Pflanzen übereinstimmen. Demzufolge sollten die Zahnmerkmale einen Hinweis auf die Umweltbedingungen liefern. Wir analysieren eine globale Erfassung von Vorkommen großer herbivorer Säuger und bioklimatischer Bedingungen. Wir identifizieren allgemeine Verbindungsmuster zwischen der Verteilung von Zahnmerkmalen und bioklimatischen Konditionen und diskutieren die Zusammenhänge. Jedes Muster kann als computergestütztes Biom angesehen werden. Unsere Analyse unterscheidet drei globale Zonen, die wir die boreal-temperierte Feuchtzone, die tropische Feuchtzone und die tropische-subtropische Trockenzone nennen. Die boreal-temperierte Feuchtzone ist hauptsächlich durch saisonale kühle Temperaturen, das Fehlen von Hypsodontie und einem hohen Anteil an Arten mit stumpfen Jochzähnen gekennzeichnet. Die tropische Feuchtzone ist hauptsächlich durch hohe Temperaturen, hohe Isothemalität, ergiebige Niederschläge und einen hohen Anteil an Arten mit eher scharfen als stumpfen Jochzähen charakterisiert. Die tropische Trockenzone schließlich zeichnet sich hauptsächlich aus durch eine ausgeprägte Saisonalität bezüglich der Temperaturen und der Niederschläge als auch durch starke Hypsodontie und Horizodontie. Wir stellen fest, dass sich die Zahnmerkmalssignatur nordafrikanischer Regenwälder deutlich von der Signatur klimatisch ähnlicher Standorte in Nordamerika und Asien unterscheidet, wo hypsodonte Arten und Arten mit stumpfen Jochzähnen zumeist fehlen. Bezüglich Kima-und Zahnsignaturen weisen die afrikanischen saisonalen Tropen viele Parallelen mit den zentralen südasiatischen Standorten auf. Interessanterweise ist das Tibetische Plateau sowohl bedeckt mit Redescriptionen aus der tropisch-subtropischen Trockengruppe als auch mit solchen aus der boreal-temperierten Feuchtgruppe, was auf eine Kombination von Merkmalen aus beiden Zonen bezüglich der Zähne und des Klimas hinweist.
Schlüsselwörter: Ökometrie; Redescription Mining; Zahnmerkmale; große Säugetiere; Data Mining
Translator: Eva Gebauer
Arabic
Translator: Ashraf M.T. Elewa
-
Review: The Princeton Field Guide to Mesozoic Sea Reptiles
The Princeton Field Guide to Mesozoic Sea Reptiles
Article number: 26.1.1R
April 2023 -