Environmental controls on the distribution and diversity of lentic Chironomidae (Insecta: Diptera) across an altitudinal gradient in tropical South America

Abstract

To predict the response of aquatic ecosystems to future global climate change, data on the ecology and distribution of keystone groups in freshwater ecosystems are needed. In contrast to mid- and high-latitude zones, such data are scarce across tropical South America (Neotropics). We present the distribution and diversity of chironomid species using surface sediments of 59 lakes from the Andes to the Amazon (0.1-17°S and 64-78°W) within the Neotropics. We assess the spatial variation in community assemblages and identify the key variables influencing the distributional patterns. The relationships between environmental variables (pH, conductivity, depth, and sediment organic content), climatic data, and chironomid assemblages were assessed using multivariate statistics (detrended correspondence analysis and canonical correspondence analysis). Climatic parameters (temperature and precipitation) were most significant in describing the variance in chironomid assemblages. Temperature and precipitation are both predicted to change under future climate change scenarios in the tropical Andes. Our findings suggest taxa of Orthocladiinae, which show a preference to cold high-elevation oligotrophic lakes, will likely see range contraction under future anthropogenic-induced climate change. Taxa abundant in areas of high precipitation, such as Micropsectra and Phaenopsectra, will likely become restricted to the inner tropical Andes, as the outer tropical Andes become drier. The sensitivity of chironomids to climate parameters makes them important bio-indicators of regional climate change in the Neotropics. Furthermore, the distribution of chironomid taxa presented here is a vital first step toward providing urgently needed autecological data for interpreting fossil chironomid records of past ecological and climate change from the tropical Andes.

Keywords: Andes; chironomids; climate change; diversity; lakes.

Figure 1

Map of South America showing the extent of the study sites. Studied lakes are in the eastern inner and outer Neotropics between 0 and 17°S and 150 and 4655 m a.s.l. White circles denote study lakes from the Andes to lowland Amazonia.

Figure 2

Subfossil larval remains of Chironomini from tropical South America. (A) Chironomini type I; (B) Chironomini type II; (C) Cladopelma type I; (D) Reithia/Pseudochironomus.

Figure 3

Subfossil larval remains of Tanytarsini from tropical South America. (A) Tanytarsus type I; (B) Tanytarsus type II; (C) Tanytarsus type III.

Figure 4

Subfossil larval remains of Orthocladiinae from tropical South America. (A) Cricotopus/Paratrichocladius type I; (B) Cricotopus/Paratrichocladius type II; (C) Cricotopus/Paratrichocladius type III; (D) Cricotopus/Paratrichocladius type IV; (E) Cricotopus/Paratrichocladius type V; (F) Cricotopus/Paratrichocladius type VI; (G) Cricotopus/Paratrichocladius type VII.

Figure 5

Subfossil larval remains of Tanypodinae from tropical South America. (A) Tanypodinae type I.

Figure 6

Chironomid assemblages for the 59 lakes. Lakes are ordered by mean annual temperature from cold to warm. Lakes are separated every 2°C along the temperature gradient with chironomid taxa shown in order of occurrence. Only taxa present in more than two lakes are shown.

Figure 7

Relative percentage of the chironomid assemblages divided by subfamily (Orthocladiinae, Tanypodinae, and Chironominae), Chironominae is further divided into two tribes, Tanytarsini and Chironomini. The lakes are grouped into elevation bins of 500 m up to lakes >4000 m a.s.l.

Figure 8

Chironomid taxon richness and evenness. (A) Rarefied taxon richness was estimated using count sizes of 23 (minimum count in the dataset). (B) Richness estimator (ACE) using rare individuals only (occurred in <10 lakes). (C) Simpson 1/D divided by the number of species was used as an independent measure of evenness to separate the signal from the richness estimates. Sample errors are 95% confidence intervals.

Figure 9

Canonical correspondence analysis (CCA) biplot of the nine explanatory variables once those with a VIF >20 were removed (i.e., Elevation and Longitude), study sites and chironomid taxa. Circles denote study sites; white triangles indicate the location of taxa. Biomes have been included as nominal variables. B1 = tropical/subtropical grass savannah‐shrubland, B2 = tropical/subtropical moist broadleaf forest, B4 = montane grass and shrub land. Only one lake occurs in B3 (tropical/subtropical dry broadleaf forest) and so its occurrence within the ordination space cannot be inferred, Biome 3 is excluded as a nominal variable.