Gregory-Wodzicki, K.M., submitted 12/98 to Paleobiology, Relationships between leaf morphology and climate, Bolivia: Implications for estimating paleoclimate from fossil floras.

Abstract.-- Fossil floras are an important source of quantitative terrestrial paleoclimate data. Many paleoclimate estimates are based on relationships observed in modern vegetation between leaf morphology and climate, such as the increase in the percentage entire-margined species with increasing temperature and the increase in leaf size with increasing precipitation. An important question is whether these observed relationships are universal or regional; for example, recent studies find significant differences between vegetation from three domains: the Northern Hemisphere, New Zealand/Australia, and subalpine zones. Also, debate exists over which statistical models of modern datasets, univariate or multivariate, provide the most accurate estimates of paleoclimate. In this study, 12 foliage samples from living Bolivian forests are analyzed using climate models based on datasets from various regions. All models produce reasonably accurate estimates of temperature and precipitation, suggesting that the climate-leaf morphology relationships for Bolivian vegetation do not differ significantly from those for Northern Hemispere vegetation. The mean leaf size for a given mean annual precipitation is smaller than for a dataset from the Western Hemisphere and Africa, but this difference is most likely due to different sampling methods. These results, along with the analysis of four other datasets, imply that the most accurate estimates for a fossil flora will be produced by the predictor dataset with the most similar climate-leaf morphology relationships. Unfortunately, our lack of understanding of why these relationships vary between the Northern Hemipshere, New Zealand/Australia, and subalpine zones makes identifying similar datasets difficult. Until we learn more, it is probably best to compare fossil floras to predictor datasets from the same leaf morphology domain. The performance of the various statistical methods depends on the nature of the predictor dataset. Multiple regression analysis tends to produce the most accurate estimates for small datasets with a narrow range of environmental variation that have similar relationships to the fossil flora, and linear regression or canonical correspondence analysis for the larger and more varied Climate-Leaf Analysis Multivariate Program dataset of J. A. Wolfe. If a similar predictor dataset is not available, then nearest neighbor analysis can still produce accurate paleoclimate estimates.

CLAMP scores for the Bolivian sites:

LMCS SJ SZ SI CN CM MA PA CB ZD SU SO TA
TLob (1) 2.3 3.4 4.9 2.6 2.6 4.1 6.3 0 7.5 1.7 0 7.8
NoT (2) 77.9 76.1 76.8 68.4 72.4 59.5 56.3 59.1 53.8 53.4 45 45.3
TRg (3) 15.1 18.2 14.6 22.4 14.5 28.4 27.3 21.6 33.1 27.6 44.2 23.4
TCl (4) 4.7 4.5 1.2 9.2 5.9 8.1 1.6 15.9 6.3 12.1 27.5 6.3
TRnd (5) 15.1 11.9 13.4 27 14.5 28.4 10.9 28.4 31.9 37.1 29.2 25
TAct (6) 7 11.9 9.8 4.6 13.2 12.2 32.8 12.5 14.4 9.5 25.8 29.7
TCmp (7) 5.8 2.3 4.9 9.2 5.3 5.4 4.7 6.8 4.4 5.2 10 3.1
SIZE (8): Nan 9.3 8 9.1 2.2 5.3 5.4 9.1 10.2 7.6 14.9 5.6 7.8
Le1 7 6.4 3.5 2.2 2.6 3.4 8.3 11.7 6.3 19.8 13.4 12.2
Le2 10 12 6.9 8.1 11.6 7.5 20.8 20.8 11.7 25.3 16.2 25.5
Mi1 18.9 21 14.6 10.7 28.7 17.7 28.2 29.9 29 26.7 20.4 37.2
Mi2 24.9 19.8 24.2 17.4 28.7 27.4 21.2 15.9 30.8 10.6 24.4 14.8
Mi3 19.1 13.4 19.5 23.1 12.3 18.6 9 11.4 9.9 1.7 11.6 1.6
Me1 8 10.6 14.8 18.5 5 12.3 1.7 0 3.6 0.9 5.4 0.8
Me2 2.1 5.4 6.7 11.1 2.4 5.2 1.7 0 0.5 0 2.3 0
Me3 0.8 3.4 0.6 6.7 3.3 2.4 0 0 0.5 0 0.7 0
AEmg (9) 37.2 13.6 22 21.1 21.1 16.2 0 9.1 5 3.4 6.7 15.6
APEX (10): ARnd 59.7 43.2 46.7 54.4 57 45.9 42.2 52.3 40 48.3 44.4 48.4
AAct 32.9 45.5 33.3 28.1 38.6 40.5 57.8 47.7 58.8 51.7 54.4 51.6
AAtn 7.4 11.4 19.9 17.5 4.4 13.5 0 0 1.3 0 1.1 0
BASE (11): BCd 24 11.4 18.7 23.7 9.2 12.2 3.1 4.5 12.5 0 3.3 0
BRnd 43.8 46.6 51.6 51.3 50 55.4 43.8 45.5 50 46.6 46.7 51.6
BAct 32.2 42 29.7 25 40.8 32.4 53.1 50 37.5 53.4 50 48.4
LW (12): LW<1:1 4.7 1.1 3.3 1.3 6.6 1.4 0 0 2.5 0 0 0
LW1-2:1 50.8 28 45.9 44.7 26.8 34 12.5 21.2 25.4 16.7 32.8 26
LW 2-3:1 30.6 31.4 36.2 38.2 43.4 37.2 45.3 37.9 30.8 30.5 37.2 28.6
LW 3-4:1 8.5 18.9 11 10.5 17.1 16.9 20.3 26.5 22.1 25.3 18.9 22.4
LW >4:1 5.4 20.5 3.7 5.3 6.1 10.6 21.9 14.4 19.2 27.6 11.1 22.9
SHAPE (13): SOb 11.6 14.4 11.8 7.9 10.5 13.5 10.9 13.6 11.3 6.9 6.7 14.1
SElp 54.7 61 58.1 59.2 61.8 62.2 65.6 50 48.8 63.8 53.3 57.8
SOv 33.7 24.6 30.1 32.9 27.6 24.3 23.4 36.4 40 29.3 40 28.1
M ln A (CLAMP) 5.67 5.82 6.21 7.08 5.85 6.33 4.92 4.64 5.37 3.86 5.28 4.43
M ln A (R-Webb) 6.14 6.3 6.57 7.21 6.2 6.66 5.37 5.34 5.62 4.45 5.51 4.68

CLAMP scores for each sample site. Numbers in Leaf margin character state (LMCS) column denote categories. Some categories have only two character states, for example 'Lobed' and 'Not Lobed' are in the category 'Lobed'. For simplicity, only one character state is usually reported. 'Teeth Acute' and 'Teeth Round' are an exception, as both are reported. Other categories, such as size, contain several character states and all are reported. Quantification of morphologic score for given leaf form: 1) if present, character state receives score of 1 divided by number of character states present for form in that category; 2) if absent, character scored 0; 3) if partly present, scored as 0.5 divided by number of character states in category. Form scores then added for each character state and divided by total number of forms to derive morphologic score. See Wolfe (1993) for details of scoring and definitions. The value for MlnA (R-Webb) was determined for each site after the equation of Wilf et al. (1998): MlnA = ai pi where ai = the means of the natural log areas of the Raunkiaer-Webb size categories and pi represents the proportion of species in each size category. MlnA(CLAMP) values were calculated using the CLAMP leaf size classification system. Because these size categories differ from the Raunkiaer-Web size categories, ai values had to be recalculated. The size of the CLAMP size categories were measured from the CLAMP size template of Herman et al. and are as follows: Na = 0-5 mm2, Le1 = 5 - 25 mm2, Le2 = 25 - 80 mm2, Mi1 = 80 - 400 mm2, Mi2 = 400 - 1400 mm2, Mi3 1400 - 3600 mm2, Me1 = 3600 - 6400 mm2, Me2 = 6400 - 10,400 mm2, Me3 10,400 + mm2. Note that these values differ from those in Forest et al. (1999); Me2 and Me3 were measured incorrectly. The measured values were used to calculate the value of a for each size category: 0.80, 2.41, 3.80, 5.19, 6.62, 7.72, 8.48, 9.01, and 9.61, with the exception that the lower size limit of Na was changed to 1 mm2 instead of 0 mm2, because the natural log of 0 cannot be calculated, and the upper size limit of Me3, which is not given in the latest CLAMP size template was set at 21,280 mm2. This value was calculated by following the general size progression of the CLAMP size categories, in which each leaf size increases its length by ~4 cm and width by ~1.8 cm, and calculating 2 size classes larger than the Me3 size category. Typically, there are very few leaves in the Me3 size category, thus the number chosen for the upper limit has little effect on the value of MlnA. Abbreviations: LMCS: Leaf morphologic character state, TLob= Teeth lobed, NoT= No teeth, TRg= Teeth regularly spaced, TCl= Teeth closely spaced, TRnd= Teeth round, TAct= Teeth Acute, TCmp= Teeth compound, Le1,2= Leptophyllous 1,2; Mi1,2,3= Microphyllous 1,2,3; Me1,2= Mesophyllous 1,2; AEmg= Apex emarginate, ARnd= Apex Round, AAct= Apex acute, AAtn= Apex attenuate, BCd= Base cordate, BRnd= Base round, BAct= Base acute, LW = Length to width ratio, Sob= Shape obovate, SElp= Shape elliptical, SOv= Shape Ovate. SJ = San José de Chiquitos, SZ = Santa Cruz, SI = San Ignacio de Velasco, CN = Concepción, CM = Camiri, MA = Monteagudo, PA = Padilla, CB = Cochabamba, ZD = Zudañez, SU = Sucre, SO = Sorata, TA = Tarabuco.

 

Climate data for Bolivian sample sites:

Station MAT WMMT CMMT GSP MMGSP 3WET 3DRY
San José 25.5 27.8 21.4 92 7.7 43.7 6.6
Santa Cruz 24.5 26.9 20.3 134 11.2 51.2 17.7
San Ignacio 24.5 26.8 20.8 122 10.2 57.6 7.9
Concepción 24.3 26.6 20.9 117 9.7 50.6 9.3
Camiri 23 26.5 17.4 87 7.3 46.6 3.1
Monteagudo 20.1 23.6 15.5 79 6.6 38.4 3.8
Padilla 17.9 20 14.9 68 5.7 39.5 1.8
Cochabamba 17.6 20.3 14.1 50 4.2 32 0.9
Zudañez 16.2 17.8 13.4 55 4.6 35.6 0.8
Sucre 15.7 17.4 12.6 65 5.4 38.4 1.2
Sorata 15.4 16.6 13.8 80 6.7 46.1 2.3
Tarabuco 12.5 13.8 10 58 4.8 36.5 1.1