Strong upslope shifts in Chimborazo's vegetation over two centuries since Humboldt

Edited by William H. Schlesinger, Cary Institute of Ecosystem Studies, Millbrook, NY, and approved August 19, 2015 (received for review May 20, 2015)
September 14, 2015
112 (41) 12741-12745
Reply to Feeley and Rehm: Land-use intensification increases risk of species losses from climate change
Naia Morueta-Holme, Kristine Engemann [...] Jens-Christian Svenning
Reply to Sklenář: Upward vegetation shifts on Chimborazo are robust
Naia Morueta-Holme, Kristine Engemann [...] Jens-Christian Svenning


Tropical regions harbor the majority of the world’s biodiversity, but there is debate about whether montane species here are able to track global warming at the same rate as in temperate regions. By following in Humboldt's footsteps and revisiting his pioneering documentation of vegetation elevation ranges, we show that the limit of plant growth has already been strongly pushed upslope. Although the rate of plant range shifts matches that found in other studies, the total magnitude of change in vegetation and glacier coverage on Chimborazo is larger than expected from warming temperatures alone.


Global climate change is driving species poleward and upward in high-latitude regions, but the extent to which the biodiverse tropics are similarly affected is poorly known due to a scarcity of historical records. In 1802, Alexander von Humboldt ascended the Chimborazo volcano in Ecuador. He recorded the distribution of plant species and vegetation zones along its slopes and in surrounding parts of the Andes. We revisited Chimborazo in 2012, precisely 210 y after Humboldt’s expedition. We documented upward shifts in the distribution of vegetation zones as well as increases in maximum elevation limits of individual plant taxa of >500 m on average. These range shifts are consistent with increased temperatures and glacier retreat on Chimborazo since Humboldt’s study. Our findings provide evidence that global warming is strongly reshaping tropical plant distributions, consistent with Humboldt’s proposal that climate is the primary control on the altitudinal distribution of vegetation.
The biological impacts of ongoing climate change (1) are already apparent in species’ poleward and upslope range shifts and earlier spring events (29). However, most studies stem from high-latitude areas and are generally restricted to dynamics across the past few decades (10). To our knowledge, only three previous resurveys have studied range shifts of tropical plant taxa, all at <4,000 m in elevation (7, 8, 11). Modeling (12) and paleoecological studies (13) suggest that tropical montane vegetation should be highly sensitive to climate change. However, researchers strongly debate whether tropical plants are tracking warming temperatures along elevation gradients, with most (although scarce) studies indicating they are lagging behind (cf. 14, 15). Such lags could have negative effects on the distributions of species dependent on certain plant taxa, e.g., as a food source (16). The question is particularly urgent given the growing evidence of systematically stronger warming rates in high-mountain environments (17).
The legacy and works of Alexander von Humboldt (1769–1859) not only constitute the foundation of biogeography, but also what is likely the oldest dataset on altitudinal ranges of plant species. The observations recorded by Humboldt and Aimé Bonpland (1773–1858) during their travels in Central and South America, and synthesized in a Tableau of Mt. Chimborazo (summit 6,268 m above sea level) and accompanying essay (18), provide a unique opportunity to study tropical vegetation changes over a period of 210 y. To our knowledge, this period is more than twice as long as any previous resurvey study based on historical biodiversity records (11, 19). We revisited the upper slopes of the Chimborazo volcano in June 2012. Our aim was to record the current elevational distribution of plants and test for upward shifts since Humboldt’s expedition, as a response to anthropogenic global warming. We sampled plant species presence and abundance along transects every 100 m of elevation between 3,800 and 5,200 m. Three main findings, comparing our surveys to Humboldt’s data, support strong upward shifts of plant distributions: a higher upper limit for plant growth, increased elevation of vegetation zones, and upward shifts in the upper range limits of most individual taxa.

Results and Discussion

The ultimate upper limit of vegetation on Chimborazo has shifted strongly upward, as expected from the warming climate. We observed seed plants growing up to an elevation of 5,185 m, constituting an upward shift of the upper vegetation limit of >500 m since 1802. According to Humboldt, the upper limit for the growth of seed plants was at 4,600 m, above which only lichens could be found (18). Our highest-elevation observations were sterile individuals of Draba aretioides Humb. & Bonpl. ex DC. and flowering individuals of Pentacalia hillii (Greenm.) Cuatrec, both found at 5,185 m.
The altitudinal distribution of vegetation zones on Chimborazo has also experienced marked shifts since Humboldt’s expedition (Fig. 1). Along Chimborazo’s elevational gradient, Humboldt recognized vegetation zones characterized by distinct species groups. Notably, he described a “Chuquiraga-gentianes-frailejón” zone with Chuquiraga, Gentianella, Gentiana, and Espeletia occurring between 2,000 and 4,100 m in 1802. At 4,100–4,600 m, he found grass-dominated “Pajonal.” We instead found the highest abundance of the Chuquiraga and gentian species in a zone between 4,200 and 4,600 m. We also found high grass abundance from 3,800–4,600 m, but with presences in plots up to 5,000 m and outside-plot observations to 5,070 m. Thus, our resurvey demonstrates an upward shift of both vegetation zones, as well as a general expansion of the Pajonal-type vegetation also at lower elevations relative to Humboldt’s observations.
Fig. 1.
Elevation changes of glacier limit and vegetation zones defined by Humboldt on Chimborazo. Current 2012 glacier limit is shown as mean ± SD. Shaded areas mark the 1802 distribution of each vegetation zone dominated by Gentianella, Gentiana, and Chuquiraga species (Gentianes); grasses (Pajonal); and vegetation in general. Curves show 2012 mean percent abundance of each vegetation zone along the elevation gradient. Abundance of all other species is shown for comparison. Overall, vegetation grows substantially above the 4,600-m limit reported by Humboldt. Although Pajonal abundance decreases drastically at >4,600 m, Gentianes species are found in abundance well above the 4,100-m historical elevation.
We also found pronounced upward range shifts in most individual taxa (Fig. 2). We found an average upward shift of 675 m at the species level (paired t test, t = 7.41; df = 50; P < 10−8), and 565 m at the genus level (t = 6.83; df = 44; P < 10−7), corresponding to an average shift per decade of 32 and 27 m, respectively. These patterns emerged consistently across multiple approaches for assembling the historical data sources, although the precise estimates of range limits varied, ranging from 586 to 787 m for species and from 557 to 700 m for genera (Fig. S1 and Tables S1S3). Although most species shifted upward, the magnitude and direction of the changes varied widely (Fig. 2 and Fig. S1), consistent with previous studies of changes in species’ elevational ranges in other regions (4, 7, 19, 20). We found some discrepancies in the accuracy of Humboldt’s elevation measurements, but no systematic bias that could explain our findings (Table S4). The results were also robust to spatial randomizations of the collection effort in 2012 (Fig. S2).
Fig. 2.
Elevation range shifts for 51 plant species over 210 y. Black, overlap between observed and historical range; green, expansion of upper range limit; red, contraction of upper or lower range limit. Dashed vertical lines show discontinuities between past and present ranges. Horizontal dashed lines mark the difference between the 1802 and 2012 vegetation limits. Species names are listed in Table S1. Historical ranges >4,600 m for species 2, 3, and 51 are shown in faded colors because they are due to the visual estimation method included when reading range limits from Humboldt’s illustrations. No plant taxa are mentioned >4,600 m in Humboldt’s texts. Note that some of the individual range shifts are driven by changes from elevations much lower than 3,800 m, likely caused by human dispersal, and are not shown in the figure.
Fig. S1.
Genus-level comparison of historical and observed elevation ranges for 45 genera. Color scheme is as in Fig. 2. Historical ranges >4,600 m for genera 43–45 are shown in faded colors because they are due to the visual estimation method. No plant taxa are mentioned at >4,600 m in Humboldt’s texts.
Table S1.
Historical elevation ranges for species
IDSpecies accepted nameElevation range, m
Narrow approachBroad approach
1Draba aretioides Humb. & Bonpl. ex DC.3,508–3,5083,508–3,5083,508–3,5083,508–3,508
2Senecio nivalis (Kunth) Cuatrec.4,483–4,8724,483–4,5613,898–4,8723,898–4,561
3Senecio canescens (Bonpl.) Cuatrec.3,703–4,6783,703–4,4833,703–4,8723,703–4,483
4Baccharis caespitosa (Ruiz & Pav.) Pers.2,339–2,9242,339–2,9241,949–2,9241,949–2,924
5Hypochoeris sessiliflora Kunth2,924–2,9242,924–2,9242,924–2,9242,924–2,924
6Lasiocephalus ovatus Schltdl.4,483–4,4834,483–4,4834,483–4,4834,483–4,483
7Xenophyllum humile (Kunth) V.A.Funk3,508–3,8983,508–3,8983,508–3,8983,508–3,598
8Chuquiraga jussieui J.F.Gmel.3,313–4,2003,500–4,2003,313–3,5472,000–4,2003,600–4,2003,313–3,547
9Valeriana alypifolia Kunth4,093–4,0934,093–4,0934,093–4,0934,093–4,093
10Azorella pedunculata (Spreng.) Mathias & Constance3,898–4,2883,898–4,0933,898–4,8723,898–4,093
11Werneria nubigena Kunth3,313–4,0933,313–3,7992,924–4,0933,313–3,799
12Astragalus geminiflorus Humb. & Bonpl.3,898–3,8983,898–3,8983,898–3,8983,898–3,898
13Gentianella cernua Kunth3,313–3,7992,924–3,3133,313–3,7992,924–3,313
14Gentiana sedifolia Kunth3,118–3,5083,118–3,5083,118–3,5083,118–3,508
15Plantago rigida Kunth3,313–3,7033,313–3,5082,924–3,8982,924–3,898
16Castilleja fissifolia L.f.2,600–3,5002,600–3,5002,600–2,6512,600–3,5002,600–3,5002,600–2,651
17Conyza cardaminifolia Kunth3,118–3,1183,118–3,1183,118–3,1183,118–3,118
18Lupinus microphyllus Desr.3,898–3,8983,898–3,8983,898–3,8983,898–3,898
19Oreomyrrhis andicola (Kunth) Endl. ex Hook. f.3,703–4,0933,703–4,0933,898–4,8723,898–4,093
20Oritrophium peruvianum (Lam.) Cuatrec.3,274–3,2743,274–3,2743,274–3,2743,274–3,274
21Perezia pungens (Bonpl.) Less.3,118–3,1183,118–3,1183,118–3,1183,118–3,118
22Cotula mexicana (DC.) Cabrera2,202–2,2022,202–2,2022,202–2,2022,202–2,202
23Eryngium humile Cav.2,631–3,8982,631–3,7032,631–3,8982,631–3,703
24Plagiocheilus peduncularis (Kunth) Wedd.3,508–3,5083,508–3,5083,508–3,5083,508–3,508
25Ranunculus praemorsus Kunth ex DC.1,949–4,0931,949–4,0931,949–4,0931,949–4,093
26Valeriana microphylla Kunth3,118–3,1183,118–3,1183,118–3,1183,118–3,118
27Baccharis arbutifolia (Lam.) Vahl3,118–3,5083,118–3,3132,924–3,8982,924–3,313
28Bartsia melampyroides (Kunth) Benth.2,631–3,1182,631–3,1182,631–3,1182,631–3,118
29Bidens andicola Kunth3,411–3,4113,411–3,4113,411–3,4113,411–3,411
30Clinopodium nubigenum (Kunth) Kuntze3,508–3,8983,508–3,8982,924–3,8982,924–3,898
31Gentianella rapunculoides (Willd. ex Schult.) J.S.Pringle3,703–4,0933,703–4,0932,924–4,0932,924–4,093
32Phyllactis rigida (Ruiz & Pav.) Pers.3,118–3,5083,118–3,5083,118–3,5083,118–3,508
33Werneria pumila Kunth3,508–3,8983,508–3,8983,508–3,8983,508–3,598
34Cerastium mollissimum Poir.2,339–2,7292,339–2,7291,949–2,9241,949–2,924
35Hydrocotyle bonplandii A.Rich.2,729–2,7292,729–2,7292,729–2,7292,729–2,729
36Lepidium bipinnatifidum Desv.2,631–2,6312,631–2,6312,631–2,6312,631–2,631
37Nertera granadensis var. granadensis1,559–1,5591,559–1,5591,559–1,5591,559–1,559
38Achyrocline alata (Kunth) DC.1,559–3,5671,559–2,436974–3,567974–2,436
39Lobelia tenera Kunth2,339–2,9242,339–2,9241,949–2,9241,949–2,924
40Acaena argentea Ruiz & Pav.2,339–3,0992,339–3,0991,949–3,0991,949–3,099
41Baccharis buxifolia (Lam.) Pers.2,534–3,1182,534–3,1182,534–3,1182,534–3,118
42Calceolaria ericoides Juss. ex Vahl3,313–3,7033,313–3,4692,924–3,8982,924–3,469
43Gynoxys buxifolia (Kunth) Cass.3,508–3,5083,508–3,5083,508–3,5083,508–3,508
44Perezia multiflora (Humb. & Bonpl.) Less.3,313–3,3133,313–3,3133,313–3,3133,313–3,313
45Azorella aretioides (Kunth) Willd. ex DC.3,703–4,0933,898–4,0933,898–4,8723,703–4,093
46Vicia andicola Kunth4,093–4,0934,093–4,0934,093–4,0934,093–4,093
47Lupinus smithianus Kunth3,898–4,2883,898–4,2883,898–4,2883,898–4,288
48Geranium humboldtii Spreng.4,093–4,4833,898–4,0933,898–4,8723,898–4,093
49Arenaria serpyllifolia L.4,093–4,4834,093–4,0933,898–4,8723,898–4,093
50Nototriche phyllanthos (Cav.) A.W. Hill2,000–4,5922,000–4,1004,093–4,5922,000–4,8722,000–4,1004,093–4,592
51Eudema nubigena Bonpl.4,093–4,7324,093–4,7323,898–4,8723,898–4,732
All ranges are given in meters and follow the narrow (NA) or broad (BA) approaches of historical data combinations described in Materials and Methods. Only taxa included in the analyses are listed. Column NA1 was used for Fig. 2. ID, species ID used in Fig. 2.
Table S2.
Historical elevation ranges for genera
IDGenus accepted nameElevation range, m
Narrow approachBroad approach
1Draba L.2,690–3,8593,508–3,8592,690–3,8593,508–3,859
2Cerastium L.2,339–4,2883,703–4,2883,703–4,2883,703–4,288
3Azorella Lam.2,100–4,1003,500–4,1002,100–4,0932,000–4,1002,000–4,1002,100–4,093
4Baccharis L.390–4,1003,000–4,100390–3,898390–4,1002,000–4,100390–3,898
5Hypochoeris L.2,924–2,9242,924–2,9242,924–2,9242,924–2,924
6Xenophyllum V.A. Funk3,508–3,8983,508–3,8983,508–3,8983,508–3,898
7Chuquiraga Juss.3,313–4,2003,500–4,2003,313–3,5472,000–4,2002,000–4,1003,313–3,547
8Valeriana L.858–4,2882,800–3,300858–4,288858–4,2882,800–3,300858–4,288
9Castilleja Mutis ex L. f.1,754–3,7031,754–3,5081,754–3,5081,754–3,508
10Lepidium L.2,631–2,6312,631–2,6312,631–2,6312,631–2,631
11Lupinus L.2,660–4,2883,500–4,1002,660–4,2882,000–4,2882,000–4,1002,660–4,288
12Luzula DC.791–4,101791–7,101791–4,101791–7,101
13Rumex L.0–31,180–3,1180–3,1180–3,118
15Werneria Kunth3,313–4,0933,313–3,7993,313–4,0933,313–3,799
16Astragalus L.2,534–3,8982,534–3,8982,534–3,8982,534–3,898
17Geranium L.2,612–4,1003,000–4,1002,612–4,0932,000–4,1002,000–4,1002,612–4,093
18Halenia Borkh.975–3,313975–3,313975–3,313975–3,313
19Plantago L.877–4,100877–3,898877–3,508877–3,508
20Conyza Less.66–4,1003,500–4,10066–3,11866–4,1002,000–4,10066–3,118
21Oreomyrrhis Endl.3,703–4,0933,703–4,0933,898–4,0933,898–4,093
22Perezia Lag.3,118–3,8403,318–3,8403,118–3,8403,118–3,840
23Eryngium L.780–3,703780–3,703780–3,703780–3,703
24Lachemilla (Focke) Rydb.400–4,3852,600–3,0002,933–4,385400–4,8722,600–3,0002,933–4,385
25Plagiocheilus Arn. Ex DC.2,670–3,5082,670–3,5082,670–3,5082,670–3,508
26Bartsia L.2,534–3,3132,534–3,3132,534–3,3132,534–3,313
27Bidens L.858–3,799858–3,411858–3,799858–3,411
28Clinopodium L.390–3,898390–3,898390–3,898390–3,898
29Sisyrinchium L.400–3,000400–3,000526–2,924400–3,000400–3,000526–2,924
30Galium L.877–2,924877–2,924877–2,924877–2,924
31Hydrocotyle L.0–3,000400–3,0000–2,7290–3,000400–3,0000–2,729
32Nertera Banks & Sol. ex Gaertn.1,559–3,3002,200–3,3001,559–3,118400–3,118400–3,0001,559–3,118
33Achyrocline (Less.) DC.2,066–3,6352,066–3,6352,066–3,6352,066–3,635
34Acaena Mutis ex L.2,339–3,3002,600–3,3002,933–3,216700–3,216700–3,0002,933–3,216
35Gynoxys Cass.2,066–3,5082,066–3,5082,066–3,5082,066–3,508
36Monnina Ruiz & Pav.780–3,508780–3,508780–3,508780–3,508
37Oxalis L.0–3,0000–3,000195–2,8850–3,0000–3,000195–2,885
38Calceolaria L.2,339–3,8982,600–3,5003,313–3,7992,534–3,8592,600–3,5003,313–3,799
39Vicia L.516–4,093516–4,093516–4,093516–4,093
40Stellaria L.2,261–4,1003,000–4,1002,261–4,0932,000–4,1002,000–4,1002,261–4,093
41Carex L.3,898–4,2883,898–4,0933,898–4,8723,898–4,093
42Nototriche Turcz.97–4,59297–4,59297–4,87297–4,592
43Eudema Bonpl.4,093–4,7324,093–4,7323,898–4,8723,898–4,732
44Ranunculus L.1,949–4,7363,898–4,2881,949–4,7363,898–4,288
All ranges are given in meters and follow the narrow (NA) or broad (BA) approaches of historical data combinations described in Materials and Methods. Only taxa included in the analyses are listed. Taxa lumped (e.g., Gentianes and Vaccinium-Gaultheria) are described in Materials and Methods. Column NA1 was used for Fig. S1. ID, genus ID used in Fig. S1.
Table S3.
Estimated upward shift of maximum range limits for species and genera 1802–2012
Taxonomic levelApproachMean upward shift, mdftP valueCI lowCI high
Abbreviations are as in Table S1. Only three taxa were reported at the species level in Humboldt’s Tableau, and therefore Approach 2 is excluded for the species taxonomic level. CI, 95% confidence interval.
Table S4.
Site elevations according to Humboldt and according to current measuring techniques for estimating accuracy of Humboldt’s elevation measurements
SiteDescriptionElevation, mSiteLatitudeLongitudeElevation, m
Santa Fe de Bogotá, ColombiaMeasurements close to Plaza Mayor (p. 218)2,662Plaza de Bolívar, Bogota DC4.59778−74.076112,616
Buga, ColombiaN-NW of Paramo de Chinche, next to Rio de las Piedras, Plaza Mayor (p. 267)973Plaza Cabal, Buga - Valle del Cauca3.89972−76.30222975
Almaguer, ColombiaMeasurements in Plaza Mayor (p. 284)2,269Parque Principal, Almaguer - Cauca1.91639−76.856392,200
Pasto, ColombiaMeasurements in Plaza Mayor (p. 286)2,616Plaza de Nariño, Pasto - Nariño1.21472−77.277782,540
Quito, EcuadorBy convento del Carmen Baxo (p. 298)2,908Monasterio del Carmen Bajo, Quito - Pichincha0.21861−78.509722,830
Cuenca, EcuadorMeasurements in Plaza Mayor (p. 367)2,633Parque Calderón, Cuenca - Azuay−2.89750−79.004172,550
Loxa, EcuadorMeasurements in Plaza Mayor (p. 369)2,064Plaza Central, Loja - Loja−3.98306−79.200002,075
Caxamarca, PeruW of Paramo d'Ianaguanga, measurements in Plaza Mayor (p. 384)2,853Plaza de Armas, Cajamarca - Cajamarca−7.15694−78.517502,730
Truxillo, PeruClose to coast, measurements in 60 toises north of Plaza Mayor (p. 386)6430 m north of Plaza de Armas, Trujillo - La Libertad−8.11028−77.1497239
Callao, Lima PeruHouse of Capitan del puerto (p. 419)6Museo Naval, Callao - Lima−12.06194−77.149728
Mexico, MexicoCapital, next to couvent de San Agustin (p. 471)2,276Ex-temple San Agustín - Cuahtémoc−19.43028−99.136672,230
Guanaxuato, MexicoHotel of Don Diego Rul (p. 511)2,083Casa del Conde Rul, Guanajuato - Pachtitlan21.01722−101.254722,020
Nevado de Toluca, MexicoFraile, highest peak (p. 524)4,623Volcán Nevado de Toluca - México19.01861−99.767504,645
Xalapa, MexicoHouse of Don Thomas Murphy next to couvent de Saint-François (p. 527)1,321Parque Juárez, Xalapa de Enríquez - Veracruz19.52639−96.923891,390
Tetimba, MexicoNext to San Nicolas de los Ranchos (p. 543)2,405Tetimpa - Puebla19.07667−98.475562,470
Cholula, MexicoCholula pyramid (p. 543)2,136Cholula Pyramid, Cholula - Puebla19.05583−98.303062,150
We assessed possible biases in the historical elevations measured by Humboldt (38) for 16 localities with sufficiently detailed descriptions to be georeferenced. Page numbers refer to those of ref. 38. Elevations measured by Humboldt originally in toises were converted to meters, multiplying by factor 1.949. Current latitude, longitude, and elevation were derived by locating the sites in Google Earth ( and GeoNames ( Although there was variation in the accuracy of Humboldt’s elevation estimates, we found no significant differences to current estimates (paired t test, t = 1.91; df = 15; P = 0.075), indicating no systematic biases in the historical measurements.
Fig. S2.
Testing sensitivity of results to collection effort. The 2012 data were subsampled to include only one or two of three plots per transect and 1/3 or 2/3 of observations (a and b, respectively). For each subsample, a paired t test was used to calculate the mean upward shift since 1802 and confidence intervals. Thick lines indicate median of mean upward shift across 1,000 randomizations of data subsampling. Whiskers show median of confidence intervals. c, mean upward shift and confidence intervals based on the full dataset. Results are shown for species and genus-level comparisons based on each of the comparison methods (NA1, BA1, etc.) described in Materials and Methods. The sensitivity analysis shows that main results are robust to collection effort and are not driven by a more intense effort in 2012 relative to Humboldt’s, which could have led to the finding of more rare or inconspicuous species.
The higher vegetation limit is consistent with the reported retreat of glaciers on Chimborazo (21), which has made new habitats available for plant colonization. Glacier coverage has retreated from 4,814 m measured by Humboldt (22) to 5,270 ± 260 m (mean ± SD) in 2013 for the glaciers on the south and east side of the volcano (21). This glacier retreat of >400 m in elevation is in line with worldwide glacier recessions linked to ongoing global warming (23). Still, recent glacier retreat may also be partially attributed to accelerated ice melting due to ash deposition from the neighboring Tungurahua volcano, active since 1999 ( Tungurahua was also active before Humboldt’s sampling of the region (1773–1776) and possibly again between 1777 and 1781 ( and ref. 24). We did not find documentation of ash emissions during the historical eruptions, but they would be expected to have had similar effects on the glaciers. Furthermore, although ash deposition may have exacerbated recent ice melting, the effects are thought to be small (21). Hence, the plant range shifts observed are most likely primarily linked to the ongoing global climate change, although not necessarily to warming alone. There is also some evidence that precipitation patterns in the area have shifted toward drier, more unpredictable regimes during the past decades, which could also contribute to glacier recession (21).
At the lower elevations studied, land-use changes likely also contribute to observed shifts (25, 26), in particular the general increase and downward expansion of grass dominance. At the lowest sites, we observed harvesting of grass tussock and signs of localized fires, both promoting Pajonal vegetation relative to Chuquiraga-gentianes-frailejón vegetation. In addition, we recorded the presence of human-dispersed nonnative species such as Trifolium repens L. and Rumex acetosella L. at 3,800–3,900 and 3,800–4,000 m elevation, respectively, which is also consistent with more intense land use in the area. Humboldt noted that there was no agriculture at >3,600 m and that it was only scattered at >3,000 m (18). In 2012, our sampling of sites at <3,800 m was hindered by strong land cover transformation to agricultural fields. Other nonclimatic anthropogenic factors such as atmospheric nitrogen deposition or other nutrients could also potentially have contributed to the observed changes in species composition and range shifts (27). However, the extent to which nitrogen deposition is driving upward range shifts in plant species is unclear, even in high-deposition regions like Europe. The increased abundance of grasses on Chimborazo could potentially be driven in part by nitrogen (28), but could also result solely from land use changes.
The general upward shift of vegetation and individual plant taxa is broadly consistent with plant species shifting their ranges to track the changing climatic conditions. National temperature records for Ecuador between 1866 and June 2012 indicate an increase in mean annual temperature of 1.46 ± 0.31 °C ( (estimate ± SE). This increase translates into an expected upshift in elevation of 243 ± 51 m, assuming a lapse rate of ∼6 °C for every 1,000 m in elevation (4, 19). Temperature measurements for Ecuador from the time of Humboldt’s expedition are not available, but the global average changed by 0.26 ± 0.45 °C between 1802 and 1866 ( Thus, following regional and global averages, the maximum expected upslope shifts would be ∼410 m since Humboldt’s study. The observed glacier retreat and vegetation responses are apparently larger than expected from regional temperature increases alone, with an observed/expected shift rate ratio of ∼1.6 for vegetation overall. The high observed shift rate likely reflects changes in precipitation regime and ash deposition from Tungurahua, as well as the impacts on vegetation from the intensified land use at lower sites (Fig. 3). Our results are thus counter to the generally lagged responses of tropical trees to climate change in shorter-term studies reporting shift rates of 0.3–0.6 (7, 14). Still, a nonnegligible proportion of our study taxa exhibited upwards shifts <410 m (Fig. 2), indicating lags. Although the upward shift we found here of 32 m per decade at the species level is faster than the reported median of 11 m across plant and animal species worldwide (29), it is similar to the 36 m per decade found for tropical plants in a recent study in Taiwan (11). Our findings of human-dispersed species shifting upward from elevations <3,800 m also strengthen concerns that the immigration of widespread generalist species may come at the cost of high-elevation endemic species and increased biotic homogenization (11).
Fig. 3.
An update of Humboldt’s Tableau. Shown is a summary of major changes in overall vegetation limit, average glacier limit, and shifts in topmost vegetation regions on Chimborazo from 1802 to 2012. The major drivers of change, climate, and land use change are represented by the bars to the right: a constant impact of climate change—in particular, increased temperature—the stronger relative impact of land use at the lower sites, mainly through intensified agriculture, and the effect of grass harvesting and local burning. Illustration of glaciers is approximate.
The substantial upward changes in the overall limit to plant growth, distribution of major vegetation zones, and upper boundaries of many species ranges show that plants are strongly responding to global warming, even within the tropics (7, 8, 16). Comparison of Humboldt’s pioneering Chimborazo vegetation survey with our resurvey thus corroborates evidence for the sensitivity of tropical montane vegetation to climate change from modeling (12) and paleoecological studies (13), highlighting that dramatic shifts can occur at relatively short, contemporary time scales. The tropical Andes constitutes a major biodiversity hotspot in terms of endemic species (30), and the high-elevation ecosystems support the livelihoods of millions of people through hydrologic control and carbon storage (31). The much-stronger climate change forecasted for the near-future (1) is therefore likely to lead to dramatic vegetation changes in tropical mountainous areas, with strong effects on the rich diversity of dependent species (16) and ecosystem services (31).

Materials and Methods


We conducted the resurvey of Mt. Chimborazo for present elevation range estimates between June 23 and July 5, 2012 on the southeastern slope of the volcano, the region originally explored by Humboldt (18). The aim was to determine the current maximum elevation of each taxon. We sampled a total of 15 transects from an elevation of 3,800 m at 100-m elevation intervals to 5,200 m, past the limit of plant growth. Each 100-m transect followed the elevation contour, and three 25-m2 quadratic plots were established at the beginning, center, and end of each transect, by using plastic poles, utility cord, and tape measures. Within each plot, we determined species presence/absence and estimated abundance as percent coverage. Coverage of <1% was set to 0.5. Additional species present along each transect were also registered, as well as species between transects. The exact elevation of species found above the 4,700-m transect was recorded to determine the limit of vascular plant presence. Elevation and geographic position were measured by using a GPS (Trimble Juno SP, ±10 m).
Sampled species were collected, photographed, and identified at the Herbario Nacional del Ecuador (QCN) and the Aarhus University Herbarium to supplement the identification in the field. Vouchers were deposited at QCN and the QCA herbarium of the Pontificia Universidad Católica del Ecuador. We sampled a total of 121 taxa (Dataset S1). We were unable to determine three of the seed species, and an additional two were determined only to family or order level (both in Asteraceae). Flowering individuals of Poaceae were determined to species level, but most sterile individuals could be determined only to family level. All Poaceae were therefore excluded from analyses, except for those regarding vegetation zones.

Historical Data Sources.

We assembled historical information on plant taxa recorded by Humboldt and Bonpland from five sources: the 1807 Essay on the Geography of Plants (18), its accompanying Tableau Physique (18, 3234), the draft to the Tableau published in 1826 (35), Humboldt’s late 1838 description of the visit to Mt. Chimborazo (22), and the descriptions of collected species in Nova Genera et Species Plantarum published 1815–1825 (36).
Essay on the Geography of Plants (18) (ref. 1 in Datasets S2 and S3): presents the detailed legend for the Tableau Physique, describing plants seen within different vegetation zones across elevations, as well as a description of the environmental variables measured (published originally in 1807).
Tableau Physique (18) (ref. 2 in Datasets S2 and S3): graphic representation of the distribution of plant taxa on Chimborazo. Some of the taxa, however, are found not on Chimborazo but in other parts of Ecuador and South America, found by Humboldt during his travels. The Tableau was originally published in 1807. We extracted its information from the Humboldt database (32, 33) and Jackson (2009) (34).
Draft to the Tableau (ref. 3 in Datasets S2 and S3): the draft includes more taxon names than the final Tableau. Like the final version, some taxa are from other regions than Chimborazo, but the figure summarizes the elevations at which Humboldt observed them (35).
Account of Chimborazo expedition (22) (ref. 4 in Datasets S2 and S3): In 1838, Humboldt presented a description of the trip to Chimborazo. The description is based on his journal written at the time of the expedition (22). Additional taxa are mentioned in this text that are otherwise omitted in the Tableau and Essay.
Nova Genera et Species Plantarum (36) (ref. 5 in Datasets S2 and S3): description of all of the taxa collected by Humboldt and Bonpland during their travels in the New World and brought back to Europe. For most taxa, a description is included of the elevation and collection locality.
For all historical data sources, we resolved abbreviated taxon names and nomenclatural updates (including synonymies) to the current Angiosperm Phylogeny Group III system using Nova Genera et Species Plantarum (36) and multiple online resources (Version 1,;; ref. 37) (Datasets S2 and S3). Historical elevations given in toises or French feet were converted to meters by using a factor of 1.949 and 3.08, respectively, following Humboldt’s own conversions (22).

Approaches to Elevation Range Comparison.

The disparate historical data sources all have advantages and disadvantages, and in some cases include inconsistent elevation estimates. Comparing the current distributions to the historical data is challenging due to the lack of a systematic design in the data collection by Humboldt. We therefore used three different approaches of data comparison to assess the sensitivity of the results to the choice of data source.

Approach A1.

Approach A1 had the broadest range of all sources. Here we used all five sources of information and assigned the broadest and most conservative elevation range reported in any source. Priority was given to the finest taxonomic resolution. Thus, a taxon could be assigned a maximum elevation higher than what Humboldt observed on Chimborazo, if it was described higher up at a different location. Figs. 13 are based on this approach.

Approach A2.

Approach A2 was a comparison with the Tableau and Essay. In this approach, we restricted the comparison with taxa specifically mentioned in the Essay or depicted on the Tableau. The rationale behind this approach is that these two sources represent Humboldt’s own final summary on Chimborazo of his observations and was validated by the botanist José Celestino Mutis (18). Most taxa were mentioned at the genus level only, so this approach could not be applied to the species level.

Approach A3.

Approach A3 was a geographic priority choice. For this approach, we chose the “best available” data source for each taxon. In cases with data from more than one source, we chose the most specific to Mt. Chimborazo, else Ecuador, else other localities in the Andes. This approach rendered the largest number of taxa for the analysis, although there is more subjectivity in the choice of source.
For images, we used visual estimation to extract information on elevation ranges (refs. 2 and 3 in Datasets S2 and S3). Because the method is more uncertain, we used both a broad and a narrow delimiting frame around the name of each taxon to record its altitudinal range. Each comparison approach is thus presented in two versions: NA1, NA2, and NA3 (narrow) and BA1, BA2, and BA3 (broad). This method allowed us to further assess the robustness of the results to the methodology.

Taxon-Level Range Shift Analysis.

We used paired t tests to compare the historical and current maximum range limit of individual plant taxa. Taxonomic resolution of the data varied, and we therefore made two comparisons for each approach: one at the species level and one at the genus level (Table S3).
For all approaches at the genus level, species were lumped to the classification used by Humboldt. Thus, Gentiana and Gentianella were lumped to “Gentianes,” Arenaria was classified as Stellaria, Phyllactis rigida was included in Valeriana, and Blumea viscosa was included in Conyza. Vaccinium and Gaultheria were also lumped (“Vaccinium-Gaultheria”), because we were unable to identify sterile individuals to the correct genus in the field. All grasses (Poaceae) were excluded from taxon-level analyses due to the difficulties in identifying sterile individuals.
Genus-level historical data were assigned by using Humboldt’s observations at the genus level where available, and otherwise assigning the highest elevation observed for species in the genus. Genera with uncertainties in the taxonomic resolution were excluded (e.g., if species in the genus now are considered to be in different genera). This exclusion was the case for, e.g., Aa, Achyrocline, Cerastium, Cotula, Hieracium, Lasiocephalus, Monticalia, Senecio, and Xenophyllum (see detailed notes in Datasets S2 and S3).
Ferns and lycophytes were excluded from the analyses, because Humboldt only noted vegetation limits for vascular plants (18).

Vegetation-Level Range Shift Analysis.

The two uppermost vegetation zones defined by Humboldt are the Pajonal, dominated by Poaceae (4,100–4,600 m), and the Gentianes region with Chuquiraga, Gentianes and Frailejón (3,000–4,100 m). To assess potential changes in the distribution of these vegetation zones, we used the abundance data recorded in the plots. For Pajonal, we summed the estimated coverage of all grasses and averaged across the three plots sampled at each elevation transect. Similarly, for the Gentianes vegetation zone, we summed the abundances of Chuquiraga, Gentiana, and Gentianella. We then plotted the mean abundances across elevation to visually assess the shifts in dominance of each vegetation zone. For comparison, we also plotted the summed abundances for all other taxa observed (Fig. 1).

Changes in Glacier Elevation.

Because glacier data for 2012 were not available, we calculated the mean of the 2013 lower limits of glaciers on Chimborazo based on data by La Frenierre (21). We assumed that any change between 2012 and 2013 would be negligible. Looking only at the glaciers on the south and east side of the volcano (glaciers Theodoro Wolf, Garcia Moreno, Boussignault, Nicolas Martin, Carlos Pinto, and Humboldt), the limit of clean ice was 5,270 ± 260 m (mean ± SD). The minimum elevation of clean ice for Chimborazo as a whole was 5,320 ± 290 m.


We thank B. Øllgaard, S. Lægaard, P. Sklenář, P. M. Jørgensen, B. Boyle, D. Fernández, and G. Peyre for assistance with validation of taxonomic determinations; Dirección Provincial del Ambiente de Chimborazo for working permits; K. Romoleroux and R. Valencia for logistic support; and F. Borchsenius for discussions on field sampling design and logistics. Assistance in the field was provided by C. Morales, guides from Andean Adventures, and park rangers at the Reserva de Producción Faunística Chimborazo. J. La Frenierre helped with data and discussions on glacier retreat. We also thank D. Ackerly, B. Blonder, R. Field, and two anonymous reviewers for useful comments that helped improve this manuscript. This work was supported by an EliteForsk Award (to N.M.-H.) and by the Ingeniør Svend G. Fiedler og Hustrus Grant (to N.M.-H. and K.E.). J.C.S. considers this study a contribution to the Danish National Research Foundation Niels Bohr professorship project Aarhus University Research on the Anthropocene (AURA).

Supporting Information

Supporting Information (PDF)
Supporting Information


; IPCC Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (Cambridge Univ Press, Cambridge, U.K., 2013).
G-R Walther, et al., Ecological responses to recent climate change. Nature 416, 389–395 (2002).
C Parmesan, G Yohe, A globally coherent fingerprint of climate change impacts across natural systems. Nature 421, 37–42 (2003).
J Lenoir, JC Gégout, PA Marquet, P de Ruffray, H Brisse, A significant upward shift in plant species optimum elevation during the 20th century. Science 320, 1768–1771 (2008).
C Parmesan, et al., Poleward shifts in geographical ranges of butterfly species associated with regional warming. Nature 399, 579–583 (1999).
RK Colwell, G Brehm, CL Cardelús, AC Gilman, JT Longino, Global warming, elevational range shifts, and lowland biotic attrition in the wet tropics. Science 322, 258–261 (2008).
KJ Feeley, et al., Upslope migration of Andean trees. J Biogeogr 38, 783–791 (2011).
KJ Feeley, J Hurtado, S Saatchi, MR Silman, DB Clark, Compositional shifts in Costa Rican forests due to climate-driven species migrations. Glob Change Biol 19, 3472–3480 (2013).
BG Freeman, AM Class Freeman, Rapid upslope shifts in New Guinean birds illustrate strong distributional responses of tropical montane species to global warming. Proc Natl Acad Sci USA 111, 4490–4494 (2014).
J Lenoir, J-C Svenning, Climate-related range shifts—a global multidimensional synthesis and new research directions. Ecography 38, 15–28 (2015).
AS Jump, T-J Huang, C-H Chou, Rapid altitudinal migration of mountain plants in Taiwan and its implications for high altitude biodiversity. Ecography 35, 204–210 (2012).
KJ Feeley, MR Silman, Land-use and climate change effects on population size and extinction risk of Andean plants. Glob Change Biol 16, 3215–3222 (2010).
MB Bush, MR Silman, DH Urrego, 48,000 years of climate and forest change in a biodiversity hot spot. Science 303, 827–829 (2004).
EM Rehm, Rates of upslope shifts for tropical species depend on life history and dispersal mode. Proc Natl Acad Sci USA 111, E1676 (2014).
BG Freeman, AM Freeman, Reply to Rehm: Why rates of upslope shifts in tropical species vary is an open question. Proc Natl Acad Sci USA 111, E1677 (2014).
I-C Chen, et al., Elevation increases in moth assemblages over 42 years on a tropical mountain. Proc Natl Acad Sci USA 106, 1479–1483 (2009).
N Pepin, et al., Elevation-dependent warming in mountain regions of the world. Nat Clim Chang 5, 424–430 (2015).
A von Humboldt, A Bonpland, Essay on the geography of plants—with a physical tableau of the equinoctial regions (1807). Essay on the Geography of Plants, ed ST Jackson (Univ of Chicago Press, Chicago), pp. 61–155 (2009).
C Moritz, et al., Impact of a century of climate change on small-mammal communities in Yosemite National Park, USA. Science 322, 261–264 (2008).
J Lenoir, et al., Going against the flow: Potential mechanisms for unexpected downslope range shifts in a warming climate. Ecography 33, 295–303 (2010).
J La Frenierre, Detecting patterns of climate change at Volcán Chimborazo, Ecuador by integrating instrumental data, public perceptions and glacier change analysis. Assessing the Hydrologic Implications of Glacier Recession and the Potential for Water Resources Vulnerability at Volcán Chimborazo, Ecuador (Ohio State Univ, Columbus, OH), pp. 64–110 (2014).
A von Humboldt Notice de Deux Tentatives d’Ascension du Chimborazo (A. Pihan de la Forest, Paris, 1838).
J Oerlemans, Extracting a climate signal from 169 glacier records. Science 308, 675–677 (2005).
ML Hall, C Robin, B Beate, P Mothes, M Monzier, Tungurahua Volcano, Ecuador: Structure, eruptive history and hazards. J Volcanol Geotherm Res 91, 1–21 (1999).
ML Forister, et al., Compounded effects of climate change and habitat alteration shift patterns of butterfly diversity. Proc Natl Acad Sci USA 107, 2088–2092 (2010).
TL Morelli, et al., Anthropogenic refugia ameliorate the severe climate-related decline of a montane mammal along its trailing edge. Proc Biol Sci 279, 4279–4286 (2012).
J-A Grytnes, et al., Identifying the driving factors behind observed elevational range shifts on European mountains. Glob Ecol Biogeogr 23, 876–884 (2014).
R Bobbink, et al., Global assessment of nitrogen deposition effects on terrestrial plant diversity: A synthesis. Ecol Appl 20, 30–59 (2010).
I-C Chen, JK Hill, R Ohlemüller, DB Roy, CD Thomas, Rapid range shifts of species associated with high levels of climate warming. Science 333, 1024–1026 (2011).
N Myers, RA Mittermeier, CG Mittermeier, GAB da Fonseca, J Kent, Biodiversity hotspots for conservation priorities. Nature 403, 853–858 (2000).
R Hofstede, et al. Los páramos andinos ¿Qué sabemos? Estado de conocimiento sobre el impacto del cambio climático en el ecosistema páramo (UICN, Quito, Ecuador), pp. 1–154 (2014).
D Doherr, The Humboldt Digital Library: Exploring innovative structures. HiN, Alexander von Humboldt im Netz (Link, Potsdam), vol VI, 35–40. (2005).
D Doherr, F Baron, Humboldt digital library and interconnectedness. Environmentalist 32, 271–277 (2011).
ST Jackson, Introduction. Essay on the Geography of Plants, ed ST Jackson (Univ of Chicago Press, Chicago), pp. 1–52 (2009).
A von Humboldt Personal Narrative of Travels to the Equinoctial Regions of the New Continent During the Years 1799-1804 (Longman, Hurst, Rees, Orme, & Brown, London) vol VI (1826).
Kunth K, von Humboldt A, Bonpland A (1815–1825) Nova Genera et Species Plantarum (Lutetiae Parisiorum, Paris), Vols 1–7.
B Boyle, et al., The Taxonomic Name Resolution Service: An online tool for automated standardization of plant names. BMC Bioinformatics 14, 16 (2013).
A von Humboldt, J Oltmanns Recueil d’Observations Astronomiques, d’Opérations Trigonométriques et de Mesures Barométriques faites pendant le cours d’un Voyage aux Régions Équinoxiales du Nouveau Continent, depuis 1799 jusqu’en 1803, par Alexandre von Humboldt; rédigées et calculées, d’après les Tables les plus Exactes, par Jabbo Oltmanns (F. Schoell, Paris) Vol 2 (1810).

Information & Authors


Published in

Go to Proceedings of the National Academy of Sciences
Go to Proceedings of the National Academy of Sciences
Proceedings of the National Academy of Sciences
Vol. 112 | No. 41
October 13, 2015
PubMed: 26371298


Submission history

Published online: September 14, 2015
Published in issue: October 13, 2015


  1. Andes
  2. climate change
  3. land use change
  4. range shifts
  5. tropical biodiversity


We thank B. Øllgaard, S. Lægaard, P. Sklenář, P. M. Jørgensen, B. Boyle, D. Fernández, and G. Peyre for assistance with validation of taxonomic determinations; Dirección Provincial del Ambiente de Chimborazo for working permits; K. Romoleroux and R. Valencia for logistic support; and F. Borchsenius for discussions on field sampling design and logistics. Assistance in the field was provided by C. Morales, guides from Andean Adventures, and park rangers at the Reserva de Producción Faunística Chimborazo. J. La Frenierre helped with data and discussions on glacier retreat. We also thank D. Ackerly, B. Blonder, R. Field, and two anonymous reviewers for useful comments that helped improve this manuscript. This work was supported by an EliteForsk Award (to N.M.-H.) and by the Ingeniør Svend G. Fiedler og Hustrus Grant (to N.M.-H. and K.E.). J.C.S. considers this study a contribution to the Danish National Research Foundation Niels Bohr professorship project Aarhus University Research on the Anthropocene (AURA).


This article is a PNAS Direct Submission.



Section for Ecoinformatics and Biodiversity, Department of Bioscience, Aarhus University, DK-8000 Aarhus C, Denmark;
Department of Integrative Biology, University of California, Berkeley, CA 94720;
Kristine Engemann
Section for Ecoinformatics and Biodiversity, Department of Bioscience, Aarhus University, DK-8000 Aarhus C, Denmark;
Pablo Sandoval-Acuña
Escuela de Ciencias Biológicas, Pontificia Universidad Católica del Ecuador, Quito, Ecuador;
Jeremy D. Jonas
Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, AZ 85721;
Department of Biology and Honors Chemistry, Tucson High Magnet School, Tucson, AZ 85705;
R. Max Segnitz
Department of Biology, Stanford University, Stanford, CA 94305
Jens-Christian Svenning
Section for Ecoinformatics and Biodiversity, Department of Bioscience, Aarhus University, DK-8000 Aarhus C, Denmark;


To whom correspondence should be addressed. Email: [email protected].
Author contributions: N.M.-H., K.E., and J.-C.S. designed research; N.M.-H., K.E., P.S.-A., and J.D.J. performed research; R.M.S. contributed new reagents/analytic tools; N.M.-H. and K.E. analyzed data; and N.M.-H., K.E., P.S.-A., J.D.J., R.M.S., and J.-C.S. wrote the paper.

Competing Interests

The authors declare no conflict of interest.

Metrics & Citations


Note: The article usage is presented with a three- to four-day delay and will update daily once available. Due to ths delay, usage data will not appear immediately following publication. Citation information is sourced from Crossref Cited-by service.

Citation statements



If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.

Cited by


    View Options

    View options

    PDF format

    Download this article as a PDF file


    Get Access

    Login options

    Check if you have access through your login credentials or your institution to get full access on this article.

    Personal login Institutional Login

    Recommend to a librarian

    Recommend PNAS to a Librarian

    Purchase options

    Purchase this article to get full access to it.

    Single Article Purchase

    Strong upslope shifts in Chimborazo's vegetation over two centuries since Humboldt
    Proceedings of the National Academy of Sciences
    • Vol. 112
    • No. 41
    • pp. 12541-12898







    Share article link

    Share on social media