Colombia), a Multiproxy Perspective

Home , Siecha Lakes, University of Pittsburgh

PALAEOENVIRONMENTS AND PALAEOCLIMATES DURING THE LATE HOLOCENE IN LAKE SISCUNSÍ (COLOMBIA), A MULTIPROXY PERSPECTIVE

A Thesis

Submitted to the Faculty of Graduate Studies and Research

In Partial Fulfillment of the Requirements

For the Degree of

Master of Science

Geology

University of Regina

Yunuén Temoltzin Loranca

Regina, Saskatchewan

December, 2018

FACULTY OF GRADUATE STUDIES AND RESEARCH

SUPERVISORY AND EXAMINING COMMITTEE

Yunuén Temoltzin Loranca, candidate for the degree of Master of Science in Geology, has presented a thesis titled, Palaeoenvironments and Palaeoclimates During The Late Holocene in Lake Siscunsi (Colombia), A Multiproxy Perspective, in an oral examination held on August 20, 2018. The following committee members have found the thesis acceptable in form and content, and that the candidate demonstrated satisfactory knowledge of the subject material.

External Examiner: Dr. David Sauchyn, Department of Geography and Environmental Studies

Co-Supervisor: Dr. Maria Velez Caicedo, Department of Geology

Co-Supervisor: Dr. Gavin Simpson, Department of Biology

Committee Member: *Dr. Tsilavo Raharimahefa, Department of Geology

Chair of Defense: Dr. Fernando Szechtman, Department of Mathematics & Statistics

*Not present at defense Abstract

This thesis provides a palaeolimnological and environmental reconstruction of the last ~2,800 years of Lake Siscunsí, located in the Eastern Cordillera of

Colombia. This study is mainly based on a high-resolution diatom record in combination with Chlorophyll 푎 and other analyses previously done in this lake, which include grain size, C/N, δ15N, δ13C, TOM, total carbonates and magnetic susceptibility.

It consists of a review of the mid and late Holocene climates in Colombia, then it summarizes the most important findings in the palaeorecords of the three

Colombian cordilleras and it briefly mentions the most significative periods of human effects. It focuses on describing the main environmental and limnological changes ocurred from ~2,854 cal yr BP until present. It describes diatom assemblage changes with the aid of the previously mentioned proxies. Diatoms were grouped by its habitat type, however the most dominant fossil diatom in this record is a new taxon, and therefore diatom guilds were used to interpret limnological changes.

Four main periods of environmental and limnological changes were recorded: (1)

From ~2,854 to 1,976 cal yr BP the lake was under low hydraulic energy, representing relatively dry conditions, with low nutrient levels and lacustrine algae as the main source of organic matter; (2) from ~1,976 to 1,572 cal yr BP the littoral area expanded under more humid conditions, with an increase of C and N, where C3 plants and macrophytes were important contributors of organic

i matter; (3) from ~1,572 to 535 cal yr BP the driest conditions in the entire record occurred along with nutrient enrichment originated by C3 plants and CAM macrophytes; (4) from ~535 to 349 cal yr BP there was a transition to humid conditions and decreased levels of nutrients under a highly disturbed environment. These conditions continued from ~349 to the present when human impact was a modifier to Lake Siscunsí dynamics.

Humid/dry episodes in the record as well as in the last decades coincide with the increases/decreases in frequency and intensity of ‘El Niño’ events in Northern

South America, being this phenomenon a possible cause of the environmental conditions recorded in the lake.

Aknowledgements

I would first like to thank my supervisors, Dr. María Vélez and Dr. Gavin

Simpson, for their continuous guidance, support, patience and motivation throughout the course of this research.

I wish to thank Katrhyn Holper, Dr. Broxton Bird and Dr. Jaime Escobar for sharing with me the sediment samples, the isotopic and geochemical data of

Lake Siscunsí. I am also indebted with Dr. Eduardo Morales for his help and guidance classifying the new diatom species, Dr. Natalia Hoyos for her time and support analyzing meteorological data, Dr. Tsilavo Raharimahefa for reviewing my thesis and Victor Vargas for his help installing the software that I used in this thesis.

I am deeply grateful to the National Council of Science and Technology of

Mexico (CONACyT) which provided me with the financial support needed to study at the University of Regina. I acknowledge the financial support from the

Faculty of Graduate Students and Research (FGSR) through the Department of

Geology in the form of scholarships. I really appreciate the teaching assistantship opportunities at the departments of Geology and International

Languages. I also thank to the SAFER project for the financial support for making possible the fieldtrip to Lake Siscunsí.

I would like to express my special appreciation to the Summer Science School for Young Scientists (TCJ-2009) of the Geosciences Centre of the National

Autonomous University of Mexico (UNAM) (Project PAPIME-PE103409) which persuaded and got me interested into the geosciences world.

I wish to thank all my friends in Canada who contributed to make Regina another home for me. iii

Table of Contents

Abstract…………………………………………………………………………………...i

Aknowledgements……………………………………………………………………...iii

Table of contents……………………………………………………………………….iv

List of figures…………………………………………………………………………...vii

List of tables……………………………………………………………………………..x

List of abbreviations and symbols…………………………………………………….xi

CHAPTER 1. Introduction ...... 1

1.1 Study area ...... 2

1.2 Colombian topography and climate ...... 6

1.2.1 Topographical characteristics ...... 6

1.2.2 Climatic characteristics ...... 7

1.3 Previous work in Lake Siscunsí ...... 9

CHAPTER 2. Holocene climates in Colombia ...... 11

2.1 Early, Middle and late Holocene climates in the Colombian Cordilleras ... 11

2.1.1 Eastern Cordillera ...... 12

2.1.2 Central cordillera ...... 15

2.1.3 Western Cordillera ...... 17

2.1.4 General correlation of palaeoclimatic records for Colombia ...... 18

2.2 Human impacts in the Colombian cordilleras ...... 22

2.2.1 Late Holocene Archaeological History in the Eastern Cordillera of

Colombia ...... 22

2.2.2 People’s arrival to the Colombian eastern highlands ...... 23

2.2.3 The pioneers of the Eastern Cordillera ...... 23

2.2.4 Cultural evidence near Lake Siscunsí ...... 26

CHAPTER 3. Methods and information for the analyses ...... 28

3.1 Coring and Sampling ...... 28

3.1.1 Age model ...... 30

3.1.2 Diatoms ...... 30

3.1.3 Grain size ...... 33

3.1.4 Magnetic Susceptibility ...... 34

3.1.5 Chlorophyll 푎 ...... 35

3.1.6 Sediment Geochemistry ...... 36

3.1.7 Loss on Ignition (LOI) ...... 40

3.1.8 Statistical analyses ...... 41

CHAPTER 4. Results ...... 44

4.1 Age model ...... 44

4.2 Diatoms ...... 46

4.2.1 Fossil record ...... 46

4.2.2 Staurosira sp. nov...... 52

4.2.3 Modern diatoms ...... 56

4.3 Grain size ...... 62

4.4 Magnetic Susceptibility ...... 65

4.5 Chlorophyll 푎 ...... 65

4.6 Sediment Geochemistry ...... 65

4.7 Loss on ignition ...... 67

4.8 Canonical Correspondence Analysis...... 71

CHAPTER 5. Discussion and concluding remarks ...... 74

5.1 Modern Diatoms ...... 74

5.2 Integrated palaeolimnological interpretation ...... 75

5.2.1 Period 1a ~2,854-1,976 cal yr BP (650-460 cm) ...... 75

5.2.2 Period 1b ~1,976-1,572 cal yr BP (460-345cm) ...... 78

5.2.3 Period 2 ~1,572-349 cal yr BP (345-75cm) ...... 81

5.2.4 Period 3 ~349- present (75-0cm) ...... 84

5.2.5 Inferred palaeoclimatic conditions ...... 88

5.3 Concluding remarks ...... 92

Bibliography……………………………………………………………………………95

Appendix A……………………………………………………………………………120

Appendix B……………………………………………………………………………150

List of figures

Figure 1.1 Lake Siscunsí in July 2017. Picture taken from the southwest facing the northeast…………………………………………………………………………….3

Figure 1.2 Lake Siscunsí in July 2017. Picture taken from the southeast facing the north………………………………………………………………………………….3

Figure 1.3 Weir in Lake Siscunsí, July 2017. Picture taken from the southeast facing the east………………………………………………………………3

Figure 1.4 (A) Location map of the Lake Siscunsí within Colombia. (B) Location of the coring site within the Lake……………………………………………………...4

Figure 1.5 Regional geology of the surroundings of Lake Siscunsí………………5

Figure 2.1 Location of Lake Siscunsí and other climate records in Colombia….14

Figure 2.2 Correlation of climate records for Colombia…………………………...21

Figure 2.3 Archaeological sites and land cultivable for maize. …………………..25

Figure 3.1 Lake Siscunsí bathymetry and modern samples……………………..30

Figure 4.1 Age depth model of Lake Siscunsí using the software Bacon………46

Figure 4.2 Lake Siscunsí accumulation rates……………………………………...46

Figure 4.3 Subfossil diatom distribution and ecology of Lake Siscunsi record…52

Figure 4.4 Staurosira sp. nov under LM…………………………………………….55

vii

Figure 4.5 Staurosira sp nov. under SEM………………………………………….56

Figure 4.6 Diatom assemblages composition in modern water samples classified according to their guild profile (high, motile or low)………………………………..61

Figure 4.7 Diatom assemblages composition of the modern grab sediment samples classified according to their guild profile (high, motile or low)…………62

Figure 4.8 Grain size ternary diagram of Lake Siscunsí core……………………64

Figure 4.9 Grain size composition of Siscunsí core plotted against time……….65

Figure 4.10 Magnetic susceptibility, TC, TN, and C/N trends……………………69

Figure 4.11 Chlorophyll 푎, δ13C , δ15N, TOM, and T Carbonates trends……...70

Figure 4.12 C/N vs δ 13C in Lake Siscunsí………………………………………..71

Figure 4.13 Results from the Canonical Correspondence Analysis……………..74

Figure 5.1 Lake Siscunsí dynamics from ~2,854 to 1,976 cal yr BP (Period

1a)………………………………………………………………………………………78

Figure 5.2 Lake Siscunsí dynamics from ~1,976 to 1,572 cal yr BP (Period

1b)………………………………………………………………………………………81

Figure 5.3 Lake Siscunsí dynamics from ~1,572 to 349 cal yr BP (Period

2)………………………………………………………………………………………..84

Figure 5.4 Lake Siscunsí dynamics from ~349 cal yr BP to present (Period

3)………………………………………………………………………………………..87

viii

Figure 5.5 Precipitation at Monguí and ‘El Niño’ years from 1962 to 1986……..91

Figure 5.6 Precipitation at Mongua and ‘El Niño’ years from 1980 to 1992…….92

List of tables

Table 3.1. Main characteristics of each diatom guild……………………………...34

Table 3.2. Environmental variables, family distributions and link functions used for fitting the statistical models……………………………………………………….44

Table 4.1. Radiocarbon ages, corresponding depth and calibrated age………..45

Table 4.2 Lake Siscunsí most dominant diatom species grouped by ecological guild according to Passy (2007)……………………………………………………..48

Table 4.3. Main characteristics of Staurosira sp. nov……………………………..54

Table 4.4 Modern water samples of Lake Siscunsí (Temperature, D.O, pH and coordinates of sampling sites)……………………………………………………….57

Table 4.5 Modern grab sediment samples of Lake Siscunsí……………………..59

Table 4.6. p- values for environmental variables after performing the CCA……72

List of abbreviations and symbols

AIC- Akaike Information Criterion

CAM- Crassulacean Acid Metabolism

CCA- Canonical Correspondence Analysis

D.O.- Dissolved Oxigen

DIC- Dissolved Inorganic Carbon

DOC- Dissolved Organic Carbon

ENSO- El Niño Southern Oscillation

GAM- Generalized Additive Model

ITZC- Intertropical Convergence Zone

LIA- Little Ice Age

LGM- Last Glacial Maximum masl- meters above sea level

SASM- South American Summer Monsoon

SEM- Scanning Electron Microscopy

SIMA- Sistema de apoyo a la toma de decisiones de la macrocuenca

Magdalena-Cauca

TC - Total Carbon

Tcarb- total Carbonates

TN - Total Nitrogen

TOM- Total Organic Matter

xii

CHAPTER 1. Introduction

Even though Colombia is one of the most studied countries in South America, from a palaeoecological and palaeoclimatic perspective, there is disagreement about which have been the most important global climatic phenomena that have affected the country during the most recent glacial and interglacial cycles. This is in part, because most climatic interpretations are based on pollen records of low resolution. Additionally, there is a knowledge gap on the effect of past climatic changes and on lakes. Therefore it is important to have more records making use of different proxies in the Eastern, Western and Central Colombian cordilleras, and thus to increase the spatial and temporal resolution of past climatic and environmental records, allowing a better understanding of the effects of climate and environmental variability on aquatic ecosystems.

One of the main reasons to study the late Holocene in the tropics is because during this period, the humans were undoubtedly a modifier of ecosystem dynamics, and also because modern climatic conditions had been established.

Through palaeolimnology, the study of past lake conditions, there is a great possibility of understanding the limnological responses to climatic variability as well as human impact in the watershed. Despite the fact that many palynologycal studies have been done in the region (e.g. Flantua et al., 2016; Bogotá.,

Hooghiemstra, & Berrio, 2016; Marchant et al., 2001; Marchant et al., 2002) they are at a low resolution scale, offering a poor possibility to find short time scale climatic interactions in neotropical regions.

The main objective of this thesis is to contribute to the understanding of late

Holocene climatic dynamics in this region and in Lake Siscunsí through a high resolution multi-proxy reconstruction (diatoms, chlorophyll 푎 and other analyses previously done in this lake, which include grain size, C/N, δ15N, δ13C, TOM, total carbonates and magnetic susceptibility).This reconstruction will provide a new perspective on the physical and biological dynamics of the lake during the last

~2,800 years.

1.1 Study area

Lake Siscunsí (Figures 1.1, 1.2 and 1.3) is located at 5° 38' 50.6034" N and -72°

47' 15.252" W on the eastern slope of the Colombian Eastern Cordillera at an altitude of 3,685 meters above sea level (masl), its maximum depth is 4.2 m

(Figure 1.4).

Ecologically, this lake belongs to the Tota-Bijagual- Mamapacha páramo complex, situated in the municipality of Sogamoso, within the Boyacá province

(Morales et al., 2007) The páramo is a high altitude ecosystem, characterized by its high capacity to collect and retain water coming from the rain and the moisture in the air, therefore is a downstream water generator. Geologically it is integrated by transitional-marine rocks from the Upper Cretaceous that include the Conejo, Ermitaño and Guadas formations composed of limestone, chert, quartzarenites, phosphatic sandstone and siltstones (Sanchez, 2005). The lake is limited to the east by the Santa María fault and to the west by the Alarcon anticline (Figure 1.5).

Figure 1.1 Lake Siscunsí in July 2017. Picture taken from the southwest facing the northeast

Figure 1.2 Lake Siscunsí in July 2017. Picture taken from the southeast facing the north

Figure 1.3 Weir in Lake Siscunsí, July 2017. Picture taken from the southeast facing the east

Figure 1.4 (A) Location map of the Lake Siscunsí within Colombia. (B) Location of the coring site within the Lake

Figure 1.5 Regional geology of the surroundings of Lake Siscunsí. Adapted from the Geological Map of Colombia (Gómez Tapias et al., 2007)

The eastern slope of the páramo complex is humid with cold and humid weather with a yearly average temperature ranging from 6° to 12°C, and a mean annual precipitation between 500 and 1000 mm, with a relative humidity index between

50 and 90% according to SIMA (Morales et al., 2007).

1.2 Colombian topography and climate

1.2.1 Topographical characteristics

Colombia is the tropical country with the largest alpine area (Veblen, Young, &

Orme, 2015) which is divided into two main regions: the Andean region, and the

Eastern plains. The first one is divided into three mountain ranges: The Western,

Central and Eastern Cordilleras (Hermelin, 2015).

 The Eastern Cordillera

Of the three mountain ranges, the eastern cordillera is the broadest and the longest one as it extends into the Mérida Cordillera in Venezuela. The highest peak (>5,700m above sea level -asl-) is located in the Sierra Nevada del Cocuy.

 The Central Cordillera

The elevation in this mountain range increases from North to South and it goes from 2,800 to 3,500m asl, and it is delimited by two rivers: Cauca and

Magdalena.

 The Western Cordillera

The highest elevation in this mountain range is between 4,000 and 5,000 m asl and it is limited on the East by the Cauca River and on the west by the Pacific

Ocean. It is covered by tropical rainforest.

The western cordillera is divided to the north in two branches, one east that goes to the Caribbean coast, and the other one, west, that converges with the Darien

Mountains in the Isthmus of Panama (Bridges, 1990; Hermelin, 2015).

All these three mountain ranges join in Southern Colombia immediately north of the border with Ecuador.

1.2.2 Climatic characteristics

Colombia is located in the northwestern part of South America, which is modulated by diverse global climatic phenomena, such as the annual migration of the Intertropical Convergence Zone (ITCZ), which defines the length and intensity of the dry and rainy seasons; the South American Summer Monsoon

(SASM) that brings changes in precipitation each year from September to April; and El Niño Southern Oscillation (ENSO) that induces warmer and drier conditions during ‘El Niño’ periods, and wetter and colder conditions during ‘La

Niña’ phase. There are also cross equatorial wind anomalies that change the levels of moisture in the air, affecting regional and local temperatures (Münnich

& Neelin, 2005). Shukla (1984) suggests that another important effect of these anomalies, is the alteration of fluvial regimes (e.g. change in flow of streams and rivers, alteration of hydrological cycle) associated with hydrologic pressures and variation in levels of soil moisture (Herrera, Sarmiento, Romero, Botero, &

Berrio, 2001). In general, in some tropical regions, especially those affected by the ITZC, annual precipitation rates tend to exceed annual evaporation rates, and this excess of moisture is transported by the trade winds; however, the frequency of precipitation in some coastal regions of South America is also associated with the topography, especially in midlands and in highlands

(Amador, Alfaro, Lizano, & Magaña, 2006).

There is another climatic phenomenon in this region, which in the tropical Andes has been called the ‘dry island´ in a ‘sea of rain’, generated by the heat produced from July to October, when large masses of moisture coming from the lowlands become trapped in the windwards of the Colombian Andes generating rain in this zones and a dry effect on the leewards of the mountains (Snow,

1976).

Münnich and Neelin (2005) state that precipitation anomalies during the rainy season (May-October) are correlated with anomalies in Sea Surface

Temperature (SST) in both, the Pacific and the Atlantic regions.

According to the Köeppen’s climate classification, Colombia has 3 main climatic zones: Equatorial fully humid in the east, equatorial monsoonal in the south central area and equatorial with a dry winter in the northern area and along the

Caribbean coast (Rubel & Kottek, 2010), with a tropical diurnal climate in the mountain regions (Kuhry, Salomons, Riezebos, & Van der Hammen, 1983).

1.3 Previous work in Lake Siscunsí

In July 2015 Broxton Bird (from Indiana-Purdue University) and his team, collected a composite core from the deepest part of lake Siscunsí (4.5m), using a modified Livingstone piston corer at 5° 38' 50.6034" N, -72° 47' 15.252" W.

Later, Holper (2018) analyzed the sand, silt, clay, lithics and magnetic susceptibility from the composite core, with the objective of reconstructing watershed erosion, runoff energy, lake levels and hydroclimatic conditions. The main aim of this study was to test the theory of the dry island effect over the last

~2500 years in Lake Siscunsí.

Holper (2018) identified 7 hydroclimatic periods in Lake Siscunsí: (1) From

~2500 to 2100 cal yrs BP there was light precipitation causing minor levels of erosion in the watershed, and low lake levels. (2) From ~2100 to 1700 cal yrs BP precipitation and erosion increased and lake levels were high. (3) From ~1700 to

1450 cal yrs BP there was an important decrease in precipitation and erosion in the watershed and therefore in lake levels. (4) From ~ 1450 to 1050 cal yrs BP there were unstable climatic conditions, but in general with high precipitation and high erosion; lake levels. (5) From ~ 1050 to 450 cal yrs BP there was a decrease in precipitation and erosion in the watershed, and low lake levels. (6)

From ~ 450 to 150 cal yrs BP and (7) from ~ 150 cal yrs BP to present, highly variable conditions under relatively dry conditions were inferred.

In order to determine if the dry island effect exists in this region, Holper (2018)

made a comparison between La Cocha (González-Carranza,

Hooghiemstra, & Vélez, 2012), Siscunsí and Ubaque (Bird et al., 2018), in

which according to their location in the mountain range, La Cocha and

Siscunsíh were taken as interior lakes, while Ubaque was considered an

exterior lake. Taking into account these considerations, Holper (2018),

suggests that the dry Island effect has been present in this region over

the last ~2,500 years, being notably stronger during the Medieval Climate

Anomaly (MCA) (~1,850-1,350 cal yrs BP) and the Little Ice Age (LIA)

(~600-100 cal yrs BP).

This meant lower lake levels at Siscunsí during the MCA, and higher lake levels during the LIA. When the atmospheric convection variability (normally caused by temperature changes) was in its extremes (LIA and MCA), then the dry island effect was stronger in Lake Siscunsí (decreases in precipitation), and when there was neutral atmospheric convection, the environmental conditions were similar in low and high elevation sites (Holper, 2018).

However, the core top samples indicate that the water level rose, but this might be not very trustful in terms of hydrological conditions because in recent times a weir was built, possibly altering the depositional settings (Holper, 2018).

CHAPTER 2. Holocene climates in Colombia

Holocene climate dynamics in northwestern South America remain unclear and are still a matter of discusssion, and in this sense, one of the main problem is that there is not a clear correlation between the local lacustrine and the palynological records as well as between the terrestrial and the marine records

(Polissar et al. 2013). This imposes a challenge when trying to correlate the local with regional changes in the past. Despite the fact that the most important variables that have influenced the Northern South America’s weather are known, a better understanding of modern climates in this region is neccesary in order to understand better the palaeoclimates.

This section is a compilation of Colombia’s Holocene climates interpreted from palaeolimnological, palaeoclimatic and palynological studies (Figure 2.1). The main aim of this section is to summarize the Holocene climate variability in the

Eastern, Central and Western cordilleras of this country. It discusses the most important palaeo records from each cordillera and then a regional correlation is attempted to produce a reference framework of Colombia’s palaeo climate.

2.1 Early, Middle and late Holocene climates in the Colombian Cordilleras

The following sub-sections deal with the main climatic reconstructions in each one of the three cordilleras.

2.1.1 Eastern Cordillera

The concave shape of the Eastern Cordillera makes it perfect to receive and retain moisture coming from the Atlantic Ocean, thus leading to high precipitation, although not as high as the Western Cordillera (Marchant et al.,

2001). This is the most studied cordillera in Colombia from the “palaeo” perspective (Figure 2.1).

Vélez et al. (2003) in Lake Fúquene recorded dry and cold conditions from

~19,700-14,200 14C yr BP, which agrees with the proposal that the driest conditions of the last 18,000 years in the Eastern Cordillera occurred during the

LGM and until 10,690 14C yr BP as reported by Behling and Hooghiemstra

(1999). At~12,635 cal yr BP, González-Carranza et al. (2012), recorded an important increase in temperature in La Cocha Lake (Figure 2.1), followed by a drop in the following 800 years.

In the Pedro Palo record (Hooghiemstra & van der Hammen, 1993; Figure 2.1), warm climatic conditions prevailed around ~11,950-11,300 cal yr BP, according to the authors, most likely related with the global Younger Dryas event. These warmer conditions were followed by a slight decrease in temperature recorded at

~10,380 cal yr BP.

Figure 2.1 Location of Lake Siscunsí and other climate records in Colombia. Pantano de Vargas (Gomez, et al., 2007), Frontino (Muñoz, et al., 2017), Fuquene (Velez, et al., 2003), Turbera de Calostros (Bosman, et al., 1994), Ubaque (Bird, et al., 2017), Patia Valley (Velez, et al., 2005)) and La Cocha (Gonzalez-Carranza, et al., 2013)

By ~9,000 cal yr BP, cool and humid conditions were reported by Bosman, et al.

(1994) in Turbera de Calostros. And Velez et al. (2003) reported that in Fuquene at ~8,680 14C yr BP there were warm and humid conditions followed by a drier phase at ~7,780 14C yr BP. These records show consistant warm and humid conditions during the Early Holocene.

From La Cocha record, it is suggested that the levels in the water column remained high from 7,000 to ~5,000 cal yr BP, most likely caused by high levels of precipitation (González-Carranza et al., 2012). On the other hand, in Turbera de Calostros (Bosman, et al.1994), detected a probable ‘regional’ dry period from ~6,650-5,200 cal yr BP with a colder interval at ~6,500 cal yr BP followed by warm conditions, coinciding with the dry period in Pantano de Vargas identified sometime between 6,000 and 5,00014C yr BP until ~2,50014C yr BP

(Gómez, Berrío, Hooghiemstra, Becerra, & Marchant, 2007). Thus the mid

Holocene seemed to have been overall humid until about 6000 yr BP where it became drier.

During the mid-late Holocene there is a controversy; Bird et al. (2017) reported a dry period from ~4,650-3,500 BP at Laguna de Ubaque, followed by an increase in precipitation from ~3,500-2,100 BP;and Vélez et al. (2003), detected a drier phase between 4,300-1,600 14C yr BP in Fuquene (located in an interior slope of the cordillera), but Bosman et al. (1994), reported cold and humid conditions from ~5,200-300 cal yr BP in Turbera de Calostros, which coincides with the humid conditions reported in the La Cocha record at ~2,550 cal yr BP as indicated by the highest levels in the water column in ~11,000 years. After

~2,100 cal yr BP, in Ubaque climatic conditions turned slightly drier due to a decrease in precipitation; in contrast, Bosman, et al. (1994) reported that humid conditions prevailed at least until ~100 cal yr BP, and that temperature was warmer from ~300-100 cal yr BP, than for the period from ~100 to 17 cal yr BP, when the dry conditions returned.

In sum, the Eastern cordillera is marked by warm and humid conditions during the late Holocene, a tendency to drier conditions during mid Holocene and highly variable conditions during the late Holocene.

2.1.2 Central cordillera

The climatic and environmental settings in the Central cordillera are more variable than in the Western and Eastern cordilleras themselves. Its location and topography make this mountain range to have great variation in temperature gradients within the year (Pabón-Caicedo, Eslava-Ramírez, & Gómez-Torres,

2001). Dueñas (1992), based on stratigraphic studies in the lower Magdalena

Basin, identified the Palaeo-ENSO phenomena for this region of Colombia, at

7,000, 5,500, 4,700, 4,000, 2,500, 2,300, 1,400, and 700 cal yr BP.

In contrast, the Cauca River record (Velez, Martínez, & Suter, 2013), recorded flooding from 4,000 to 3,260 cal yr BP, and a subsequent decline in the hydraulic energy, that could be related with a decrease in precipitation. Wijmstra and Van

Der Hammen (1966) proposed that between 3,700 and 3,500 cal yr BP wet conditions prevailed in this region, and in addition, Marchant and Hooghiemstra

(2004), detected a regional shift in vegetation indicative of moister conditions

15 between ~4,000 and ~3,500 yr BP. After this period, Velez et al. (2013) suggested that dry conditions prevailed between ~2,800 and ~1,300 cal yr BP whereas Garcia et al. (2011) suggested that fire events were rare and wet conditions prevailed at ~2,200 cal yr BP in the same region.

Giraldo-Giraldo, et al. (2017), suggested for the region nearby the Ruiz Volcano

(El Triunfo record) that at ~1,700 cal yr BP there was an increase in precipitation under warmer climates, then, by ~1,550 cal yr BP there is a shift to drier conditions and by ~1,400 yr BP they reported that wet conditions returned, which matches with the conditions recorded by Velez et al. (2013) in which after ~1,350 cal yr BP and until ~850 cal yr BP, there is a notable change in the hydraulic regime to increasing flood events in the Cauca region.

The El Triunfo record (Giraldo-Giraldo et al., 2017) shows also that at ~1,170 cal yr BP there is a warm and dry episode but after ~1,300 cal yr BP wet conditions returned but at this time accompanied by a cold climate, with a shift to warm and dry conditions from ~1,200-1,010 cal yr BP which, according to them, is probably part of the effect of the medieval warm period.

After this time, the el Triunfo wetland (Giraldo-Giraldo et al., 2017) showed cold conditions at ~650, ~580, ~520, and ~490 cal yr BP. The event recorded at ~650 cal yr BP is also reported in the Frontino record (Muñoz et al., 2017) and

Giraldo-Giraldo et al. (2017) suggests that this cold climate was caused by the

Wolf Solar Minimum event. They also mentioned that the coldest and wettest event occurred sometime between ~280-230 cal yr BP.

Records from the Central cordillera are scarce and they mainly describe the mid and late Holocene, however they showed that for most part, this mountain range have prevailed under wet conditions with two drier episodes.

2.1.3 Western Cordillera

The Western Cordillera is mainly affected by the Pacific Ocean and by the climate phenomena occurring in the south of the Equator. This is the least studied cordillera with only one study in the Cordillera itself, the Frontino record

(Muñoz et al., 2017) and one from the lowlands, in the Patia Valley (Vélez, et al.,

2005).

The Frontino record shows that between 12,000 and 9,700 cal yr BP climatic conditions were cold and wet, with a shift to warmer temperatures at ~9,700 cal yr BP, probably reflecting the Holocene Thermal Maximum. These conditions were followed by the wettest period in this region recorded between 7,500 and

4,200 cal yr BP (Muñoz et al., 2017).

In the lowlands of El Patía region, Vélez et al. (2005) suggest that the climatic conditions from 8,350 to 7,690 cal yr BP were mainly dry, and then from 7,690 to

6,690 cal yr BP climates were highly seasonal and gradually changing, from wet to dry. From 6,690 to 3,890 cal yr BP, conditions were mainly dry.

In the Amazonia, Liu and Colinvaux (1985), suggested that a dry period occurred between 4,200 and 3,150 yr BP in the Amazonia; and at ~900 yr

BP.These dry conditions were caused by severe ENSO events in coastal Peru environments (Marchant et al., 2001).

Since few records are available and seemed contradictory (El Patia and

Frontino), a general correlation is not attempted.

2.1.4 General correlation of palaeoclimatic records for Colombia

In this section, the most complete records from Colombia are considered, and then a regional climate comparison is attempted (Figure 2.2). The records selected were: Frontino (Western cordillera); Fuquene, Ubaque and La Cocha

(Eastern Cordillera). These records were selected because of their length, resolution (Ubaque and La Cocha), and because they include changes in the water levels identified through either diatoms, grain size or lithics, except the record from Frontino, which is only based on palynological investigations.

For the purpose of this section, the climatic conditions are going to be described in two periods: (1) The early Holocene from ~10,000-5,000 cal yr BP and (2) the mid and late Holocene from ~5,000-0 cal yr BP.

 Early Holocene (from ~10,000 to ~5,000 cal yr BP)

The Frontino record (Muñoz, et al., 2017) shows a declining trend in rainfall with warm conditions from ~11,410 to 7,500 cal yr BP, and a tendency to a drier environment, with a very dry and cold period between ~7,700 and ~7,500 cal yr

BP, coinciding with the dry conditions observed in Fuquene by Velez, et al.

(2003) which lasted until ~8,70014C yr BP. Thus it seems that these dry conditions were regional and based on these two records, temperature in the

Western Cordillera was warm, while cold in the eastern cordillera.

These dry conditions are also observed at La Cocha from ~9,000-6,000 cal yr

BP (González-Carranza et al., 2012), and in the lowlands of the Patía region

(Velez, et al., 2005). Unlike these records, Fuquene (Velez, et al., 2003) in the

Eastern Cordillera does not present this drier climate for this time span (Figure

2.1).

The period from ~7,500 until ~4,200 cal yr BP was the wettest period detected in the Western Cordillera with 3 periods of further increases in precipitation at: (1)

~7,500-7,000 cal yr BP, (2) ~5,600-5,300 cal yr BP and (3) at ~4,200cal yr BP

(Muñoz, et al., 2017). This is also observed at La Cocha (González-Carranza, et al., 2012) from ~7,000 to ~5,000 cal yr BP, but not in Fuquene, where drier conditions were recorded (Velez, et al., 2003).

Dry and warm conditions seemed to have prevailed in the Western Cordillera as indicated by the Frontino record (Muñoz, et al., 2017), and in the eastern

Cordillera by La Cocha record (González-Carranza et al., 2012).

 Mid and Late Holocene (from ~5,000 to present)

The Frontino record (Muñoz, et al., 2017) shows an important cold period between ~6,000 and ~4,000 cal yr BP and the maximum depth in the water column between ~5,000 and ~4,200 cal yr BP, most likely caused by a significant increase in precipitation.

La Cocha Frontino Fúquene Ubaque (G onz alez - (Muñoz, et al., 2017) (Velez, et al., 2003) (Bird, et al., 2017) Carranza, et al., 2013)

10,000

9,500

9,000

8,500

8,000

7,500

7,000

6,500 Decreasing 6,000 water Significant levels 5,500 cooling 5,000 Increased max. activity of 4,500 Water ENSO depth 4,000 Reduced 3,500 precipitation Decrease 3,000 in rainfall

2,500 Decreased

2,000 wet/dry/wet precipitation and 1,500 water levels

1,000

500

Wet Cold Rainfall maxima Humid Warm Precipitation

Dry

Figure 2.2 Correlation of climate records for Colombia. Climatic conditions obtained from Frontino (Muñoz, et al., 2017), Fuquene (Velez, et al., 2003), Ubaque (Bird, et al., 2017), and La Cocha (Gonzalez-Carranza, et al., 2013)

On the contrary, the Fuquene record (Velez, et al., 2003), shows dry conditions from ~7,000- 1,60014C ± 60 yr BP, coinciding also with the dry conditions detected at Ubaque (Bird, et al. 2017) from ~4,650-3,500 cal yr BP. This period was followed by a rapid decrease in rainfall at Frontino (Muñoz, et al., 2017) from ~4,200-2,600 cal yr BP and a warming period from 4,000-2,500 cal yr BP with wetter periods at ~2500 and ~500 cal yr BP. Wet conditions between

~3,500 – 2,500 cal yr BP were recorded in both, La Cocha (González-Carranza, et al., 2012) and Ubaque (Bird, et al., 2017), however, this is contrary to the dry conditions observed at Fuquene (Velez, et al., 2003).

The Frontino record (Muñoz, et al., 2017) shows that after ~2,400 cal yr BP there was a rapid decline in temperature, but a rise in temperature occurred from ~200 cal yr BP until present, and pararell to that, wet conditions prevailed along these periods. La Cocha (González-Carranza, et al., 2012) and Ubaque (Bird, et al.,

2017) records, also shows wet conditions after ~2,400 cal yr BP but Fuquene presented dry and cold conditions (Velez, et al., 2003).

In sum, during the Holocene it is hard to find a climatic pattern that is consistently shown in the Colombian cordillera records, because the climate dynamics seem to be very different between the three of them, in long and short scale periods. In those terms Mayewski et al. (2004), state that climate variability in short time periods has been highly dynamic during the Holocene and it is still poorly understood, that is why it is important to explain changes in short time spans to have a better understanding of human and ecosystem dynamics.

The records previously mentioned do not show a clear correlation pattern and thus, the global climatic episodes described by Mayewski et al. (2004) are difficult to identify, particularly in the late Holocene. It is possible that the geographical and geomorphological settings of the records described above, generate micro-climates. Also, as stated by Mayewski et al. (2004), the anthropogenic impacts become more evident during the late Holocene, making the interpetation of the records more difficult (due to secular trends in climate, e.g. global warming).

2.2 Human impacts in the Colombian cordilleras

2.2.1 Late Holocene Archaeological History in the Eastern Cordillera of Colombia

Much has been said about the impact that recent society is causing to the planet and to the future climate patterns, however we haven’t understood the impacts of the change in land use that the first settlers caused, not only to the landscape but also to the biogeochemical cycles, to the flora-fauna species interactions and as a consequence to the climate. Hodell et al. (2000) suggest that the disturbances caused in the past, can be the same ones that we have in the present, but in an accentuated way.

Human activities have caused different levels of impact on a given area, depending on many factors such as: population size, proximity to a settlement, water availability, etc. This is the reason why it is important to understand long term human-nature interactions in areas where palaeoecological studies are being done.

The main objective of this review is to describe the main environmental changes as a result of human activities in the Eastern Cordillera and thus help explain the ecological changes obtained from the diatom and Chl 푎 analysis from the core of

Lake Siscunsí. The following section briefly describes how the first settlers arrived in Colombia, then it continues by presenting the archaeological evidence in the Eastern Cordillera and finally it discusses the most important activities in each one of the different archeological stages.

2.2.2 People’s arrival to the Colombian eastern highlands

The first settlers of the America´s travelled all the way from the renowned

Beringia towards the South of the continent, as suggested by Gruhn (1978). This migration was done all along the coast with the intention of having two ways of obtaining their food, hunting and fishing. Once the early populations crossed

Panama, only few groups ventured inland and by ~12,000 yr BP people started to settle in the Colombian lowlands according to Lavallée (2000). Most of the people who settled in Colombia did it in the high mountain forests, the ‘Sabana de Bogotá’, the Calima, Cauca and Magdalena valleys, using the Magdalena and Cauca river valleys as a connection between the coast and the inlands

(Dillehay, 2000).

2.2.3 The pioneers of the Eastern Cordillera

Approximately 25 archaeological sites have been identified in Colombia, however only a few of them are located within the eastern cordillera:

Tequendama, El Abra, Nemocón, Sueva, Tibitó, Guatavita, Tunja and

Sogamoso (Figure 2.3)

Figure 2.3 Archaeological sites and land cultivable for maize. Adapted from the Atlas of ancient America (Coe et al., 1986)

According to Bruhns (1994), the first four sites mentioned above demonstrated evidence of being specialized hunters of deer, peccary, agouti, and guinea pigs until the abandonment of this places at around 4,000 BC. Dillehay, (2000) reports that near from Tequendama site (Figure 2.3), some stone tools made of animal bones were found, indicating that the time span between 12,000 and

10,100 cal yr BP was still a hunting period; he also comments that the site Tibitó

(Figure 2.3) which dates from approximately 11,740 cal yr BP, worked as an important food processor centre or a ‘hunting station’. Nemocón which is believed to be ~8,000-9,000 years old, served as important site of salt extraction.

People were dedicated to trade this good with other settlements around the

Eastern and the Central cordilleras (Bruhns, 1994). The date of Sueva is a matter of debate, Correal (1990) dated the site between 13,000- 10,000 cal yr

BP meanwhile Dillehay (2000) thinks that the age correspond between 11,000 and 10,000 cal yr BP. At El Abra, which dates from ~12,500 cal yr BP, some evidence of cultivation of yam and pumpkin was found after ~2,795 cal yr BP, it is also suggested that their main agricultural method was the use of elevated terraces (Correal-Urrego, 1990).

Hunting was the main source of food; between 9,500-8,000 cal yr BP the most consumed animal was the guinea pig. However, in Colombia the guinea pig was not domesticated until ~ 2,500 cal yr BP. In terms of agriculture, the first evidence of Zea maize was found by Piperno (1989) at around 7,000 cal yr BP using pollen and phytoliths from a site in the Calima valley; this suggests that at some time after this date, they began an omnivorous diet.

2.2.4 Cultural evidence near Lake Siscunsí

In the eastern cordillera, the most proximal archeological site to Lake Siscunsí is the Sogamoso site (Figure 2.3). According to Rodriguez (2007), the first settlers of this place used the Chicamocha River valley as the main route to get there.

In this site, four main periods were identified by Fajardo-Bernal (2016), the

Herrera period (400 BC-800 AD), the Early Muisca period (800-1,200), the Late

Muisca (AD1,200-1,537 AD), and the Colonial period (1,537-1,810 AD).

The Herrera period (400 BC-800 AD) consisted in the initial settling; at this time, around 50-100 people inhabited the Sogamoso valley and as a “new technology” they began to cultivate in a smaller scale, another main activity was pottery

(Fajardo-Bernal, 2016).

During The Early Muisca Period (1,200-1,537 AD), population grew around 10 times what it was during the Herrera period, and as a consequence the agriculture increased. In the course of the Late Muisca period, however, population decreased ~17% due to different local and/or regional ‘forces’, and after that, some evidence suggest that they probably began to adjust their crops according to the type of soil (Fajardo-Bernal, 2016).

In the Colonial Period (1,537-1,810 AD), agricultural practices were more effective and intense, and when the Spaniards arrived in Colombia, a great exploitation of goods in the highlands started, however the number of inhabitants at this time is still unknown, probably because the building of new cities wiped out all evidence (Beltrán-Beltrán, 2008).

The agricultural and hunting practices were intense in the Eastern Cordillera since 400 BC, these activities can have an effect on the microflora an in general in the limnological conditions of water bodies due to the transport of some particles of sediment and nutrients.These anthropogenic effects impose a challenge when trying to correlate palaeoclimatic record, mainly because it is hard to disentangle the human factor from the ecosystem natural dynamics, which has a different intensity in each watershed. These effects can be triggered by population growth or the development of the agriculture, and can turn into changes in biological community structures, eutrophication, and dissapereance or substitution of some biological species (Zhang & Mei, 1996).

CHAPTER 3. Methods and information for the analyses

3.1 Coring and Sampling

The core was collected in 2015 by Dr. Bird and his team, and it was stored at

4°C in the Palaeoclimatology and Sedimentology Laboratory at the Indiana

University-Purdue University Indianapolis (IUPUI).The sediment core was extracted from the deepest part of the lake with a modified Livingstone corer.

Seven sediment samples from the sediment-surface interface (grab samples,

GS) were collected from different parts of the lake with an Ekman Grab Sampler

(Figure 3.1).

In 2017, samples approximately every 5 cm from the original core were sent to the University of Regina. 125 samples were analyzed for diatoms (119 from the core and 6 from the water-sediment interface).

During the summer of 2017, four modern water samples (M1-4) from different parts of the littoral zone were collected in the Lake in order to know the modern diatom flora of the lake, as well as the modern pH, Oxygen and Temperature.

Figure 3.1 Lake Siscunsí bathymetry and modern samples. “GS” represent the grab sediment samples and “M” represent the water samples. Modified and adapted from Holper (2018)

3.1.1 Age model

Holper (2018) collected charcoal samples from the core at seven different depths (50 cm, 133.5 cm, 198.5 cm, 206.5 cm, 314 cm, 478 cm, 579.5 cm), and pre-treated the samples with HCl and C14H12O8 following the procedures of

Abbott and Stafford (1996). After this pre-treatment, samples were sent to the

Keck Carbon Cycle AMS Facility at the University of California Irvine for 14C dating analysis.

An age model was produced with Bacon v.2.2 (Blaauw & Christen, 2011), making use of the IntCal13 calibration curve with R version 3.4.4 (R Core Team,

2018) as an interface. Accumulation rates were estimated at 5 cm resolution intervals.

3.1.2 Diatoms

Diatoms were prepared according to the methodology proposed by Battarbee,

(1986), with some adjustments, as follows:

1. Diatoms were highly abundant in the sediment, so only 0.02 g of dry

sediment were taken from the samples.

2. Samples were put in beakers with 30 ml of H2O2, then placed on a hot

plate and heated until the reaction was completed.

3. Samples were left to settle 24 hours in the fume hood, and after that time,

the supernatant was siphoned off and replaced with distilled water, the

samples were stirred; this procedure was carried out twice with the

purpose of removing residual H2O2 from the samples.

4. Once again the samples were stirred and then 0.3 ml aliquots were taken

and pipetted onto glass slides.

5. The glass slides were left in the fume hood until the aliquot dried out.

6. Glass cover slides were mounted into the glass slides with Zrax® (RI ~

1.7+).

119 slides were prepared and analyzed under a 1000× oil immersion lens (N.A

=1.5) on a Light microscope OLYMPUS CX41. At least 400 diatom valves were counted per sample.

Tilia version 2.0.41 (Grimm, 2015) was used in order to plot the stratigraphic variations of the diatoms, and diatom abundance was expressed as percentage of the sum of total diatoms.

A cluster analysis using a constrained incremental sum of squares (CONISS) was made to aid the identification of the diatom assemblage zones.

Diatom slides for modern water samples were prepared at the University of the

Andes, in Bogotá, Colombia. Conductivity, pH and D.O (Dissolved Oxygen) were measured in situ at the same sites of diatoms collection with an EcoSense

DO200A Dissolved Oxygen Meter and a Milwaukee Mi 105 Portable pH/Temperature Meter respectively.

The limnological reconstruction was made based on the autecology of the most representative taxa (see Apendix B), and by classifying all diatom species according to its ecological guild.

Root (1967) defines “guilds” as a group of organisms that benefit from the same kind of nutrients or resources in a comparable manner independently to its taxonomic position. Sommer et al. (2012) suggest that these guilds can be analyzed as functional components of the biological community.

In terms of diatom guilds, Passy (2007) designated three ecological guilds for diatoms: low, high and motile profile, where the low profile includes prostrate, adnate and erect species which move slowly and are quite short in comparison with other species; the high profile encloses chain-forming, erect, stalked and centric species; and finally, the motile profile consists of fast moving species.

The main characteristics of each guild are described in Table 3.1.

Table 3.1. Main characteristics of each diatom guild. Adapted from Passy (2007), Gottschalk & Kahlert (2012), Stenger-Kovács, Lengyel, Crossetti, Üveges, & Padisák (2013)

Low profile High profile Motile profile

 It is resistant to high current  It is dominant in low current  It is a nutrient competitor in velocities velocities nutrient rich environments  It is dominant in low nutrient  It favours the macrophytes  It tolerates pollution environments growing  It shows ease of adaptation  It shows strong response to  It is sensitive to flow  It does not have an nutrient enrichment disturbance analogue of C-S-R  Macrophytes are excellent  It holds C and R strategies strategies habitat for low profile  It is proportional to nutrient  It has strong correlation with species abundance inorganic TN  It Persist when nutrients are  It can inhabit in the biofilm  It is sensitive to conductivity high  It is dominant in high  It shows a multi-habitat  It is adapted to high resource periods behaviour disturbance  It requires high luminosity  It is associated to cold water  It holds R strategies  It is dominant in low temperature  It is dominant under alkaline dissolved oxygen  It is highly adapted to water conditions environments discharges  It is tolerant to low light  It has high affinity to epilithic intensities, but it prefers high habitats luminosity conditions  It decreases in periods of It is sensitive to grazing recurrent floods  It increases when TN decreases  It shows a multi-habitat behaviour.

3.1.3 Grain size

Grain size distribution in water bodies is representative of hydraulic transport or depositional processes, helping with water level reconstructions (Sun et al.,

2002).

Holper (2018), measured grain size of Lake Siscunsí core every 0.5 cm by using a Malvern Mastersizer 2000; values were calculated as the average of three replicates measurements. Grain size results were parsed into 49 grain size bins between 0.2 and 2,000 µm.

For the interpretation of this proxy in this thesis, higher proportions of sand were taken as indicative of expansion of the littoral zone (that is, the littoral zone getting closer to the core), or more energetic hydraulic conditions (e.g. flood) as more energy is needed to transport heavier sediments. Lower sand proportions were taken as contractions of the littoral zone (that is, the littoral zone is moving away of the core), and more pelagic conditions are stablished as less energy is needed to transport lighter sediments like silt and clay (Flügel, 2004).

3.1.4 Magnetic Susceptibility

Magnetic Susceptibility measures the magnetic mineral content which is associated with erosional processes in lake sediments (Murdock, Wilkie, &

Brown, 2013). Higher magnetic susceptibility indicates increasing catchment inputs, and peaks can reflect increased inwash as a result of flood pulses

(Caballero et al., 2002). In shallow lakes, sediment mixing is not significant enough to alter the chronological control of the record (Larsen & MacDonald,

1993).

Magnetic Susceptibility of Lake Siscunsí core was measured by Holper (2018), using a Multi-Sensor Core Logger (MSCL) at the University of Indiana. Samples

34 were taken at room temperature every 0.5 cm, the measurements were reported in SI units (×10-5).

3.1.5 Chlorophyll 푎

Pigments in water bodies can mainly arise from in situ production, this is, plant material and or resuspension. These pigments are involved in different biological processes of various organisms until they are bleached, oxidized or destroyed, and its degradation occurs just before depositing to be part of the fossil record

(Leavitt, 1993).

Out of the four different groups of chlorophyll, this study is only centered in Chl 푎 and its derivatives (due to economical limitations), which are indicative of total algal and plant abundance in lakes, and thus a proxy for primary productivity.

Chlorophyll 푎 is not an independent indicator, and in this sense, Michelutti et al.

(2010) have reported a positive correlation between nitrogen, phosphate and carbon with Chl 푎, indicating that these trends in elements might explain changes in trophic levels. Furthermore, Douglas, Smol, & Blake (2000), have reported high levels of Chl 푎 in lakes that have been used for sewage waste.

For Lake Siscunsí core, 89 sediment samples were analyzed every 5 cm with some intervals every 10cm (due to the scarcity of sediment for some depths) at the Palaeoecological Environmental Assessment and Research Lab (PEARL) at

Queen’s University. Spectrally-determined Chlorophyll 푎 concentrations were obtained following the methodology proposed by Michelutti et al. (2010).

The absorbance peak area was recorded between 650 and 700 nm through a visible reflectance spectroscopy (VRS), and then the following formula was applied to get the Chlorophyll 푎 + derivatives:

푪풉풍풐풓풐풑풉풚풍풍 풂 + 풅풆풓풊풗풂풕풊풗풆풔 = 0.0919 × 푝푒푎푘 푎푟푒푎650−700푛푚 + 0.0011

(Michelutti et al., 2010)

3.1.6 Sediment Geochemistry

 Stable isotopes

δ 13C

δ 13C in water bodies is often used to track and identify the main sources of organic matter (Fogel & Cifuentes, 1993). In general, during periods of great productivity in lakes (increased biomass), the content of 13C in the water rises, having heavier δ13C signatures as a consequence (Coletta, Pentecost, & Spiro,

2001). Several studies (Coleman & Fry, 1991; J. E. Keeley & Sandquist, 1992;

Meyers & Teranes, 2001; Rau, 1978) have shown that the values of C4 plants range from -17‰ to -9‰, C3 plants from -32‰ to -20‰, macrophytes from -

50‰ to -11‰, and phytoplankton from -47‰ to -26‰. Within the group of macrophytes, some species have shown CAM (Crassulacean Acid Metabolism) physiotypes, in which as a mechanism of adaptation under stressing conditions, they can rather opt to concentrate inorganic carbon or to use its acid metabolism

36 in order to save water (Luttge, 2010). This is a common group of plants in the high-altitude paramo biomes (Luttge, 2010) and they can have either a facultative CAM character or an intermediate C3/CAM mechanism to resist environmental changing conditions. In the tropics, Boom et al. (2014) have reported δ13C signatures ranging between -20 ‰ and -27 ‰ for the first group and between -20‰ and -24‰ for the second group.

According to Meyers & Teranes (2001) plants can metabolize C, getting high levels of 13C, then heavier δ13C as a consequence (up to -9 ‰); in contrast, when the water level increases there is a high influx of CO2 , causing low levels of δ13C (Andrews, Riding, & Dennis, 1993).

δ15N

Generally, δ15N in surface and ground water ranges between 0‰ and 10‰

(Kendall, 1998), when its value goes from -4‰ to 4‰, it reflects natural sources, and nitrate and/or ammonium from fertilizers and between 8‰ and 18‰, it is interpreted as derived from sewage waste (Heaton, 1986).

Heavier signatures of δ15N can be sometimes correlated with ammonia volatilization (ammonium loss under alkaline conditions) or denitrification

(normally produced by anoxic conditions) (Leng et al., 2005). In contrast, lighter signatures of δ15N are normally associated with rises in ammonium concentration.

For instance, lakes with high levels of organic matter (OM) have reported values of δ15N > +10 (Leng & Marshall, 2004). Values of δ15N between -5 to +18, have

37 been interpreted as originated from terrestrial plants, nitrogen fixing terrestrial plants have values of -6 to +6, whereas microbial nitrogen fixation or terrestrial plant debris in suspended material have values of -2 to +4 (Fogel & Cifuentes,

1993).

 Elemental C, N and C/N ratio

The major N sources that can have an effect in lakes include: soil plants, aquatic macrophytes, rain and allochtonous and autochthonous organic matter (Talbot,

2001). This element can be present in several ways e.g. ammonia, nitrates, and nitric acid; from these three, the first one is the main detritus produced by plants and animals (Maitland, 1990). N has a strong direct effect on bacterial growth

(Farjalla, Esteves, Bozelli, & Roland, 2002) and increases in algal productivity are known to increase nitrogen values in lake sediments (Robinson, 2001).

In lakes, the main inputs of carbon are surface water, groundwater, atmospheric processes and regolith (Myrbo, 2012). Once it enters the lake, carbon is transformed by both, abiotic and biotic processes where the main products are

Dissolved Inorganic Carbon (DIC), Dissolved Organic Carbon (DOC), and particulate inorganic carbon. In tropical lakes the main form is DOC (Tranvik et al., 2009). Carbon in lake sediments is normally accumulated as carbonates and as organic matter; this two products can add up to the total proportion, however detrital mineral debris can comprise an important percentage (Myrbo, 2012).

C/N ratio has been used to identify possible sources of organic matter in lakes depending on the value of this ratio. In general this ratio is indicative of the

38 sources of organic matter in lake sediments. C/N ratios of 5 or 6, indicate presence of phytoplankton and zooplankton (Bordovskiy, 1965). Organic matter which is derived primarily from algae and phytoplankton normally ranges between 4 and 10 (Meyers, 1994).

When this ratio increases, and has values above 20, it has been interpreted as indicative of terrestrial organic matter from the catchment arriving to the lake. In contrast, when this value decreases, it can be interpreted as indicative that algal productivity is the major source of OM (Mahapatra, Chanakya, & Ramachandra,

2011), suggesting increases of the lake productivity in situ (Thevenon, Adatte,

Spangenberg, & Anselmetti, 2012), whilst higher C/N values (10–20) might indicate a mix of aquatic and terrestrial organic material (Mackie, Leng, Lloyd, &

Arrowsmith, 2005; Zong, Lloyd, Leng, Yim, & Huang, 2006).

154 sediment samples every 5 cm were sent to the Light Stable Isotope Mass

Spec Lab in the Department of Geological Sciences at the University of Florida, where sediment samples were freeze-dried and crushed with a mortar and pestle prior to geochemical analysis.

Total carbon (TC wt %) and total nitrogen (TN wt %) were measured using a

Carlo Erba NA1500 CNS elemental analyzer. Samples for carbon stable isotope analysis of organic matter were treated with 2N HCl to remove carbonate and then washed with distilled water to remove chloride. Approximately 50 mg of carbonate-free bulk sediment was loaded into tin sample capsules and placed in a 50-position automated carousel on the elemental analyzer.

Non-acidified samples were used for nitrogen isotopes. Combustion gases were carried in a helium stream through a Conflo II interface to a Finnigan-MAT 252 isotope ratio mass spectrometer. All carbon isotope results are expressed in standard δ notation relative to VPDB. Nitrogen isotopes are expressed as the per mil deviation from air.

3.1.7 Loss on Ignition (LOI)

 Total Orgaic Matter (TOM)

The main sources of C in lake sediments are related with organic matter produced from atmospheric CO2, C3 land plants, phytoplankton (C3 algae), terrestrial plants in the watershed, algal organic matter (Meyers & Lallier-Vergès,

1999)

 Total carbonates

Morris et al. (2013) suggested that the content of carbonates in lake sediments can be controlled by 2 main sources, water evaporation and dissolution of bedrock: decreases in carbonates might indicate increases in the water column.

In order to measure LOI in Lake Siscunsí, Holper (2018) collected 1cm3 of sediment sample each 2 centimeters throughout the 650cm core, and calculated the weight differences starting ignition after 550°C and 1000°C respectively, by following a modified methodology from Dean (1974) and Heiri et al. (2001) in order to measure the Total Organic Matter (TOM %) and Total Carbonates

(Tcarb %).

3.1.8 Statistical analyses

All the data collected was divided into 2 subsets. The first one contained the diatom species composition, and the second one contained the “environmental variables”: Chlorophyll 푎, N, C, TOM, Tcarb, δ15N, δ13C, C/N and magnetic susceptibility (MS).

The analyses done included:

 Canonical Correspondence Analysis

In order to measure the linear relationships between diatom species and environmental variables such as chlorophyll 푎, δ15N, δ13C, C/N and magnetic susceptibility, a Canonical Correspondence Analysis (CCA) was performed using the ‘vegan’ package (Oksanen et al., 2018) in R version 3.4.4. The CCA was tested using Monte Carlo simulations with 999 permutations, and the environmental variables included in the CCA were statistically significant when p<0.05.

 Trend analyses

In order to estimate trends in the time series of the environmental parameters,

Generalized Additive Models (GAMs) were fitted due to its flexibility to fit non- linear relationships in the predictor variables by adjusting the degrees of freedom for the cubic splines smoothing (Clark, 2013; Matthiopoulos, 2011; Wood, 2006).

According to the properties of the data, a distribution family and a link function were selected for each variable (Table 3.2) with the purpose of establishing ‘a

41 relationship between the mean of the response variable and a smoothed function of the explanatory variable’ (Guisan, Edwards, & Hastie, 2002). Then, a

95% confidence interval was calculated for each trend in each environmental parameter.

For sand, silt and clay, Dirichlet Regression for compositional data was used with the aid of the ‘DirichletReg’ package (Maier, 2013) for R. A smoothed function of time was used to estimate trends in the relative composition. The identification of an appropriate level of smoothing was done using the Akaike

Information Criterion (AIC, Akaike, 1973).

Table 3.2. Environmental variables, family distributions and link functions used for fitting the statistical models

Variable Distribution Link function family

C/N Gamma logarithmic

δ13C Gaussian identity

δ15N Gaussian identity

Chlorophyll 푎 Gamma logarithmic

TC Gamma logarithmic

TN Gamma logarithmic

TOM Beta logistic

TCarb Beta logistic

Magnetic Susceptibility Gaussian identity

CHAPTER 4. Results

4.1 Age model

The age model for Lake Siscunsí was built through Bacon v.2.2 (Blaauw &

Christen, 2011) using the 4 radiocarbon dates provided by Holper (2018) (Table

4).There is a lack of chronological control at the uppermost part of the core, since the shallowest sample sent to date was located at 198.5 cm.

Table 4.1. Radiocarbon ages, corresponding depth and calibrated age

Composite Cal yrs Average 14C age (BP) ± BP Depth (cm)

199.5 1310 120 1055

207.5 1180 25 1089.3

479 2020 40 2035.9

580.5 2420 45 2494

The calibrated model indicates that the record is ~ 2854 years old (Figure 4.1) and that the average accumulation rate is ~0.5 cm/yr (Figure 4.2), however in some depths the model indicated rates as low as 0.40 cm/yr and as high as 0.65 cm/yr. Accumulation rates were plotted in greyscale indicating a 90% confidence interval (grey dashed lines), the mean value was marked with a dashed red line.

Figure 4.1 Age depth model of Lake Siscunsí using the software Bacon. (a) Markov Chain Monte Carlo model iterations. (b) Prior (green) and posterior (gray) distribution of accumulation rate. (c) Prior (green) and posterior (gray) distribution of the age model. (d) Calibrated 14C dates and age depth model. The central red line indicates the ‘best’ fit for the model based on the weighted mean. The gray shade indicates the 95% confidence intervals.

Figure 4.2 Lake Siscunsí accumulation rates. The mean value is marked with red, the grey lines represent a 90% confidence interval and the darker shades show the precision of the measurements.

4.2 Diatoms

4.2.1 Fossil record

A total of 49 taxa were identified in the record, belonging to 31 genera. The most diverse genus was Stauroneis (with 6 species), followed by Navicula and

Gomphonema (4 species each) and Staurosira and Pinnularia (3 species each) as observed in Figure 4.3.

The dominant diatom species is a new taxon which belongs to the Staurosira genus according to Eduardo Morales, the world’s specialist in this genus. An article is in preparation with the aim of presenting it to the scientific community.

In this thesis, I refer to it as Staurosira sp nov. A brief description of this taxon can be found in section 4.2.2.

Diatoms were classified according to its habit into planktic, benthic and tychoplanktonic, and taking into account the diatom guild classification suggested by Passy (2007). In 3 out of the 4 diatom zones found in the Lake

Siscunsí record, the most predominant diatom guild was the high profile and according to the habitat type the dominance was marked by benthic species.

According to those criteria, the most dominant species of the Lake Siscunsí core were grouped as indicated in Table 5.2.

Table 4.2 Lake Siscunsí most dominant diatom species grouped by ecological guild according to Passy (2007)

Low profile High profile Motile profile

 Cymbella aspera  Staurosira construens  Navicula viridula  Amphora libyca  Staurosira sp. nov  Navicula trivialis  Cocconeis placentula  Fragilaria tenera  Navicula placentula  Rhoicosphenia abbreviata  Pseudostaurosira  Sellaphora pupula  Epithemia turgida pseudoconstruens  Epithemia adnata  Achnantes subhudsonis var.  Gomphonema truncatum  Stauroneis anceps krauseli  Gomphonema affine  Encyonema silesiacum  Pinnularia gibba  Staurosira pinnata  Pinnularia acrosphaeria  Tabellaria floculossa  Punctastriata mimetica  Stauroneis acidoclinata  Planothidium  Stauroneis phoenicenteron frequentissimum  Stauroneis gracilis

Three main zones were identified after running the CONISS analysis (Figure

4.3). Zone 1 was subdivided into 2 subzones (1a, 1b). Relative abundances are presented in parentheses.

 Zone 1a (650-460cm; ~2854-1976 cal yr BP)

This zone comprises mainly benthic diatoms such as Staurosira sp. nov. (10-

90%) and Staurosira construens, (5-50%) with an isolated peak of 85% at 500 cm. The presence of other species like Staurosira pinnata and Cocconeis placentula represent not more than 20% each. Epitemia adnata, Gomphonema affine, Rhoicosphenia abbreviata, Epithemia turgida and Planothidium frequentissimum are present but with relative low abundances (<10%).

This zone is dominated by the high profile guild, which oscillated between 60 % and 100% of representability and the rest of the proportion (up to 30%) was comprised by low profile diatoms; the presence of motile diatoms was almost null (Figure 4.3).

 Zone 1b (460-345cm; ~1976-1572 cal yr BP)

In the same way as in the previous zone, the diatom relative abundance is marked by Staurosira sp. nov. and Staurosira construens, however there is an important decrease in the abundance of S. sp.nov. from 430 to 350 cm and a simultaneous increase in S. construens and Cocconeis placentula (20-45%).

Epithemia turgida, Sellaphora pupula, Gomphonema affine, Rhoicosphenia abbreviata, Epithemia adnata and Staurosira pinnata have the same relative abundance values as in the previous zone (less than 10%). Species like

Stauroneis acidoclinata and Diploneis smithii var. dilatata make their first appearance in this zone (not more than 5% of relative abundance).

In this zone the percentage of low profile diatoms increased (up to 50%) compared with the previous zone, however the most predominant guild was the high profile which reached up to 80% of representativeness (Figure 4.3).

 Zone 2 (345-75cm; ~1572-349cal yr BP)

Staurosira sp.nov. increases again in this zone, fluctuating between 40 and 85%.

Stauroneis anceps appears for first time at 340 cm and keep its abundance (1-

15%) up to Zone 3. The abundance of Sellaphora pupula and Pinnularia acrosphaeria increases (1-20% and 1-7% respectively) concurrently with the one

48 of Stauroneis acidoclinata from approximately 160 cm up to 110 cm. Cocconeis placentula decreases significantly in comparison to the previous zones; along this zone its maximum range goes only up to 7%.

The diatom dominance was marked by the high profile guild (from 70 to 100%), followed by the motile profile which had a relative increase compared with the previous zones (reaching up to 25%). The low profile diatoms decreased and their representativeness were kept under 10% in this zone (Figure 4.3).

 Zone 3 (75-0cm; ~349-cal yr BP to present)

In this zone, it is important to highlight the appearance of different diatom taxa that were not recorded in previously. This zone is dominated initially by

Sellaphora pupula (20-50%) between 75-45 cm, but starting approximately at 45 cm Cocconeis placentula becomes dominant (10-50%), followed by Fragilaria tenera from 60 to 0 cm (unique appearance in this zone (5-40%) and

Planothidium frequentisimum from 70 to 0 cm (up to 30%). Gomphonema truncatum, Navicula viridula, Navicula trivialis, Stauroneis phoenicenteron and

Gomphonema acuminatum appear for the first time along the core with less than

10% of relative abundance. Stauroneis acidoclinata peaks from 75-45 cm (from

5-18%). The abundance of Epithemia adnata, Gomponema affine and

Encyonema silesiacum in this zone is the greatest of all zones along the core

(up to 20%).

In this zone there was a marked change in the diatom guilds dominance (Figure

4.3), the high profile guild decreased significantly, while the motile and low

49 profile diatom guilds increased. At the beginning of this zone, the motile profile reached its maximum between 75 - 50 cm (50%), then from 50-0 cm the low profile guild was the most dominant (up to 60%).

Figure 4.3 Subfossil diatom distribution and ecology of Lake Siscunsi record. Diatoms were classified by habitat type: Planktic, benthic, tycho-planktic, and benthic-planktic for the ones that can live in both habitats. They were also classified by profile: Low (gold), high (gray) and motile (blue). Staurosira sp. nov. was not included in the sum by habitat (due to its unknown behaviour)

4.2.2 Staurosira sp. nov.

A new taxon belonging to the Staurosira genus was found in Lake Siscunsí sediment core. The highest abundance of Staurosira sp. nov. is between 90 and 350 cm depth with isolated peaks from 463 cm to 479 cm.

This taxon presents a rectangular frustule shape (Figure 4.4), is generally found forming chainlike colonies (Figure 4.4 (2,3,7)) thanks to the presence of linking spines which are clavate with insertions to the mantle. The striae frequency is from 14 to 15 in 10 µm. It varies in length from 11.0 to 45.0 µm and in width from 3.5 to 5.0 µm.

The valve view under Light Microscopy (LM) of Staurosira sp. nov can be easily mistaken with Pseudostaurosira alvareziae, Pseudostaurosira americana or

Pseudostaurosira subsalina. It is important to be careful when in doubt to classify this kind of diatom genus, there is no way to differenciate this taxon to the others previously mentioned by using LM, however this species can be easily identified under Scanning

Electron Microscopy (SEM) (Figure 4.5), and this lead us to a genus level where the difference between the species previously mentioned and Staurosira sp. nov is that these species have the spines on the virgae, and not on the vimines as the rest of the taxa belonging to Pseudostaurosira (E. Morales, personal communication, January 7,

2018; Morales, 2001).

The main characteristics for its identification are shown in Table 4.3.

Table 4.3. Main characteristics of Staurosira sp. nov

Staurosira sp. nov.

Frustules shape (Girdle view) Rectangular

Chainlike colonies Yes

Way of forming chains Spines

Linking Spines shape Clavate with insertions to the mantle

Spines location On the virgae at the valve face margin

Valves Isopolar, slightly elliptical to lanceolate

Apices Broadly rounded

Striae frequency 14-15/10

Punctate, uniseriate, alternate, parallel at

Striae description the central area and slightly radiate

towards the apices

Length 11.0-45.0

Width 3.5-5.0

Figure 4.4 Staurosira sp. nov under LM

Figure 4.5 Staurosira sp nov. under SEM

4.2.3 Modern diatoms

 Water samples

In all of the 4 modern water samples analysed, 37 taxa were identified, belonging to 22 genera. The most diverse genus was Gomphonema (with 4 species), followed by

Navicula (with 3 species).

Temperature, pH, D.O and the coordinates of modern water samples are listed in Table

4.4.

Table 4.4 Modern water samples of Lake Siscunsí (Temperature, D.O, pH and coordinates of sampling sites)

Temperature D.O Sample pH Coordinates (°C) (mg/L) M1 7.78 9.50 6.61 5°38'49.83"N, 72°47'20.93"W M2 7.70 12.80 6.90 5°38'45.46"N, 72°47'20.66"W M3 8.20 11.40 7.48 5°38'45.85"N, 72°47'17.60"W M4 7.00 13.50 5.70 5°38'55.26"N, 72°47'8.43"W

o Modern Water Sample 1 (M1)

This sample was taken at 5°38'49.83"N, 72°47'20.93"W which corresponds to a delta formed in the lake (Figure 3.1). The most dominant species were Eunotia biggiba and

Planothidium frequentissimum with more than 40% of representativeness and 20% respectively, followed by Synedra ulna, Epithemia adnata, Frustulia vulgaris, Navicula trivialis and Pseudostaurosira pseudoconstruens with 10% of abundance. Encyonema silesiacum, Navicula radiosa, Sellaphora pupula, Asterionella formosa, Gomphonema accuminatum, Gomphonema affine and Gomphonema augur represented approximately 5% of the abundance and the other species had less than 2% of representativeness. In this sample the diatom guild that was best represented was the 56 motile profile having 6 species with more than 10% of abundance, followed by high profile species (Figure 4.6).

o Modern Water Sample 2 (M2)

This sample was taken from the littoral zone in front of the delta at 5°38'45.46"N,

72°47'20.66"W (Figure 3.1), the most dominant species was Gomphonema truncatum with more than 40% of representativeness, followed by Pseudostaurosira pseudoconstruens with 25% of abundance. Pinnularia subcapitata, Pinnularia acrosphaeria, Navicula trivialis, Planothidium frequentissimum and Meridion circulare had slightly more than 10% of abundance. The other species have less than 5 % of abundance. The high profile guild was the most represented with 2 species above 20%, followed by the low profile guild (Figure 4.6).

o Modern Water Sample 3 (M3)

This water sample was taken in the littoral zone in between water sample 2 and the weir at 5°38'45.85"N, 72°47'17.60"W. The most dominant species was Eunotia biggiba with close to 40% of representativeness, followed by Gomphonema truncatum with 25%.

Stenopterobia delicatissima and Cocconeis placentula and Tabelaria floculossa were close to a 10% of representativeness. The high profile diatom guild was the best represented with 2 species above 20% of representativeness followed by the motile profile guild (Figure 4.6).

o Modern Water Sample 4 (M4)

This water sample was taken in the littoral zone close to an inflow of the lake at

5°38'55.26"N, 72°47'8.43"W. The dominance was marked by 3 species: Tabelaria floculossa, Pinnularia acrosphaeria, Cocconeis placentula and Eunotia biggiba, in which the first one of them had above 40% of abundance. The low profile guild was the best represented followed by the motile profile guild (Figure 4.6).

 Grab sediment samples

In all of the 6 modern grab sediment samples, a total number of 26 taxa were identified, which belong to 19 genera. The most diverse genera were Navicula and Stauroneis

(with 3 species each), followed by Epithemia, Gomphonema and Pinnularia (with 2 species each).

These grab sediment samples were classified by Holper (2018) as transitional, littoral or profundal zones according to its location within the lake (Figure 3.1). Coordinates and their corresponding zone are listed in Table 4.4.

Table 4.5 Modern grab sediment samples of Lake Siscunsí (Coordinates of sampling sites). Adapted from Holper (2018). Transitional indicates a changing zone between the littoral and the profundal zone or viceversa; Profundal indicates the deepest part of the lake; and littoral represents the littoral zones.

Sample Coordinates Zone GS1 5°38'51.98"N, 72°47'12.82"W Transitional GS2 5°38'49.74"N, 72°47'19.86"W Littoral GS3 5°38'49.38"N, 72°47'18.20"W Transitional GS4 5°38'48.95"N, 72°47'15.86"W Profundal GS5 5°38'50.86"N, 72°47'13.09"W Transitional GS6 5°38'52.62"N, 72°47'10.93"W Transitional

o Transitional zones

. Grab samples 1,3, 5 and 6

The most dominant diatoms in these samples are Gomphonema truncatum and

Navicula trivialis, followed by Pseudostaurosira pseudoconstruens with close to 20% of representativeness in the three different samples of the transitional zone. Stauroneis exigua has an abundance greater than 15% in 3 out of the 4 samples. Epithemia turgida is also present close to a 10% of abundance. The most dominant diatom guild was the motile profile followed by the high profile diatoms (Figure 4.7).

o Littoral zone

. Grab sample 2

This sample was marked by the dominance of Pinnularia subcapitata with close to 40% of abundance, followed by Tabellaria floculossa with more than 12%. Navicula trivialis,

Epithemia turgida and Cocconeis placentula were represented approximately 10% of abundance. The low profile diatom guild was the best represented in this zone with 3 species above the 10% of abundance, followed by the motile profile (Figure 4.7).

o Profundal zone

. Grab sample 4

This sample is dominated by the presence of Navicula trivialis with more than 25% of abundance, followed by Epithemia turgida with 20% of abundance. And with close to a

15% of abundance, Gomphonema truncatum and Pseudostaurosira pseudoconstruens were also present. Diatoms with motile profile dominated in this zone with one species above 25% and 3 above 10% of representativeness, followed by the high profile diatoms (Figure 4.7).

Figure 4.6 Diatom assemblages composition in modern water samples classified according to their guild profile (high, motile or low)

Figure 4.7 Diatom assemblages composition of the modern grab sediment samples classified according to their guild profile (high, motile or low)

In the following sections, the other analyses will be discussed based on the fossil diatom zones described in section 4.2.1

4.3 Grain size

Along the whole core, grain size is distributed between silty loam and loamy sand

(Figure 4.8). Clay oscillated between 0% and 30% during the whole record (Figure 4.9), while silt was ranging between 20% and 70% and sand had the greatest variability of all these 3 sizes between 0% and 80%. A general opposite trend is observed between sand and silt and a similar trend is observed in clay and silt (in different proportions).

In Zone 1a (Figure 4.9) from 650 up to 545 cm the sediment was composed of ~ 75% sand, 20% silt and 5% clay. From 545 until ~450 cm the sediment size completely changed, being ~60% silt, 20% sand and 20% clay. In zone 1b sand ranged between

50% and 70%, silt decreased again and represented ~30% of the sediment composition, while clay remained very stable at ~10%.

In Zone 2 (Figure 4.9) sand levels decreased down to 40% at ~220cm; silt content increased up to 50% and clay kept its value in ~10%. At ~160 cm there is another change, silt went down to 20% and sand rose up to 80% again, while clay dropped to

5%. By the end of zone 2 and the start of zone 3 sand and silt started to change their proportions and at ~20cm sand reached the lowest value in this zone (25%), silt reached 50% and clay registered an abundance of ~20%.

Figure 4.8 Grain size ternary diagram of Lake Siscunsí core. Dots and colour represent the composition of each sediment sample. Clay (blue), silt (green), sand (red).

Figure 4.9 Grain size composition of Siscunsí core plotted against time. Coloured intervals represent diatom zones.

4.4 Magnetic Susceptibility

Since little variation was observed in this record, it will not be reported according to the diatom zones. Magnetic Susceptibility ranged from -1.25 to -0.25 with only 2 significant changes at ~ 545 cm and 160 cm (Figure 4.10). From the deepest part of the core until

~545 cm there was a decreasing trend from -0.25 to -0.80 SI, after which values remained around -0.80 SI until the middle of zone 2 (~178 cm) where it starts a climbing trend until the top of the core.

4.5 Chlorophyll 푎

Chlorophyll 푎 values ranged from 0.01 up to 0.07 mg/g with three significant changes at

~306, ~108 and ~55 cm (Figure 4.11). In zone 1a and 1b values were stable and oscillated around 0.02 mg/g, the highest values occurred at ~544 and ~475 cm

(0.025mg/g in both). The transition between zone 1b and zone 2 (from ~ 363 to ~306 cm) is characterized by a marked increasing trend in Chl 푎 values, where the highest recorded value was 0.070, after which Chl 푎 values were oscillating between 0.047 and

0.070 mg/g. Between 108 and 55 cm there was a significant decrease from 0.05 mg/g to 0.03 mg/g.

4.6 Sediment Geochemistry

 Elemental C, N and C/N

Carbon (wt %) oscillated between 12% and 19% with five significant changes at ~626,

~544, ~420, ~306 and ~178 cm (Figure 4.10). The start of zone 1a was characterized by a dropping trend from 18% to 13% going towards 544cm, after which C increased up

65 to 18.5% in zone 1b. Above this depth until 178 cm, it went down to 15% and remained around that value in zone 3.

Nitrogen (wt %) ranged between 1.2% and 2.1% with 4 significant changes at ~544,

~449, ~249 and ~178 cm (Figure 4.10). From the start of the core to 589 cm there was a decreasing trend going from 1.6% to 1.3% after which N percentages went up to 1.7% towards ~420 cm where values remained around 1.7% until ~306 cm when the tendency increased up to 2.1% having the highest value at 178 cm. Above this depth, N kept a decreasing trend.

C/N ratio oscillated around 9 and 13 with five significant changes at ~450, 363, 108 and

19 cm (Figures 4.10 and 4.12). The start of zone 1a is characterized by a decreasing trend between 650 cm and 544 cm going down from 12 to 9.5. In zone 1b there is an increase from 9.5 up to 11, and at 391 cm C/N decreases to 9.5. In zone 2, C/N was stable, it remained around 9.5. In the transition between zone 2 and 3, from ~108 to

~73cm C/N went up to 11, and above 19 cm it increased.

 δ15N

This variable ranged from 1 to 3 with 4 significant changes at ~ 565, ~365, ~110 and

~55 cm (Figure 4.11). In zone 1a δ15N increased until the end of zone 2 (~355 cm), it went up from 1 to 2.3 and remained stable until ~110 cm where it increased reaching the highest value (3 ‰) at ~55 cm, above that depth and going towards zone 3, δ15N started to decrease.

 δ13C

δ13C ranged from -27 to -21, it showed 3 main significant changes at ~390, ~305 and

~125 cm (Figures 4.11 and 4.12). From the bottom of the core up to ~390 cm (zone 1a and 1b) values oscillated between -27 and -25, above this depth values peaked to -22 towards ~305 cm and remained between -23 and -22 until ~125 cm when a decreasing trend started getting down to -27 in the uppermost part of the core.

4.7 Loss on ignition

Total carbonates (Tcarb) were very low, they oscillated between 2% and 5% (Figure

4.11). From the deepest part of the core up to 545 cm values remained around ~4%, above this depth there was a decreasing trend up to the end of zone 1b (~335 cm) where values went down to 3%. From the start of zone 2 values went up to 4% towards

~200 cm after which they decrease getting down to 2.5% in the surface.

TOM was very high compared with the Tcarb proportion, percentages moved between

25% and 40% and it had 3 main changes along the record at ~565, ~335, ~220 cm

(Figure 4.11). From the bottom of the core to ~565 cm there was a decreasing trend going from 36% to 26%, after which TOM rose up to 37% at 420 cm. Above that depth until the end of zone 1b (~355 cm) values got down to 30%. For most of the zone 2 and in zone 3 values fluctuated between 30 and 35%.

Figure 4.10 Magnetic susceptibility, TC, TN, and C/N trends. Points represent the samples, the grey shadow represent a 95% confidence interval for the plotted trends. Coloured areas represent diatom zones.

Figure 4.11 Chlorophyll 푎, δ13C , δ15N, TOM, and T Carbonates trends. Points represent the samples, the grey shadow represent a 95% confidence interval for the plotted trends. Coloured areas represent diatom zones.

Figure 4.12 C/N vs δ 13C in Lake Siscunsí. Coloured areas represent diatom zones and textured areas represent the type of OM producers. Points represent each sediment sample and its respective depth. For lacustrine algae and C3 plants the reference values were taken from Meyers and Terranes (2001), and for the C3/CAM plants the reference values were taken from Boom et al. (2014).

4.8 Canonical Correspondence Analysis

The CCA showed that the first 2 axes explained 52.37% (eigenvalue on axis 1 (λ1) =

0.325, and eigenvalue on axis 2 (λ2) =0.198) of the variability in the diatom species and the environmental data. δ13C was the longest vector, followed by Chl 푎, C/N ratio,

Magnetic Susceptibility and δ15N (Figure 4.12).

Chlorophyll 푎, δ15N, δ13C, C/N ratio and magnetic susceptibility are environmental variables that seem to explain significantly (p<0.05) the diatom distribution in the core from Lake Siscunsí (Table 4.5). Additionally, the Monte Carlo test results showed that of the previously mentioned variables (with more than 95% of confidence), C/N ratio, Chl

푎, δ 13C and δ 15N were the principal factors affecting subfossil diatoms, while Magnetic

Susceptibility had the lowest contribution explaining the diatoms distribution in the lake compared to the other variables (Table 4.5).

Table 4.6. p- values for environmental variables after performing the CCA

Environmental δ13C Chl 푎 C/N ratio Magnetic δ15N variables Susceptibility

p - value 0.001 0.001 0.001 0.023 0.001

CCA axis 1 is positively associated with Chl 푎 and δ13C, but negatively associated with

Magnetic Susceptibility and C/N. CCA axis 2 is negatively associated with all the environmental variables (Figure 4.13).

Three main groups of species can be identified in the CCA, these groups are associated with the highest abundances of diatom species per diatom zone. It can be observed that groups 1, 2 and 3 contain species from diatom zones 1a, 1b, 2 and 3 respectively

(Figure 4.12).

Also, in the CCA it can be seen that in Lake Siscunsí record, Magnetic Susceptibility and C/N are highly correlated and at the same time they are opposite to the signal of

Chl 푎 and δ13C (these last two variables being correlated).

Eunotia implicata, Craticula ambigua, Staurosira sp. nov., Stauroneis anceps and

Stauroneis neohyalina showed a strong relationship with δ13C. Caloneis lewiisi,

Cymbella aspera, Diploneis smithii, Navicula placentula, Epithemia turgida and

Staurosira construens were inversely related to Chlorophyll 푎.

Gomphonema affine, Stauroneis gracilis, Epithemia adnata, Achnantes subhodsonis var. kraeuselli, Sellaphora pupula, Pinnularia subcapitata and Gomphonema augur demonstrated a close relationship with δ 15N. In contrast, Punctastriata mimetica was inversely related to δ 15N.

Rhoicosphenia abbreviata, Frustulia vulgaris, Navicula viridula, Stenopterobia delicatissima, Navicula trivialis and Fragilaria tenera were highly correlated with

Magnetic Susceptibility and C/N.

Figure 4.13 Results from the Canonical Correspondence Analysis. Diatom species, Magnetic Susceptibility, C/N, Chlorophyll 풂, δ 13C and δ 15N. Groups 1, 2 and 3 contain species that had their highest peak in abundance in diatom zones 1a, 1b, 2 and 3 respectively. The labels in red show the species that had more than 5% of representativeness, whilst the labels in yellow show the species that had less than 5% of representativeness.

CHAPTER 5. Discussion and concluding remarks

5.1 Modern Diatoms

 Littoral zone water samples

The littoral diatom assemblage (Figure 4.6) indicated fresh-brackish waters with species highly tolerant to organic pollution like Pseudostaurosira pseudoconstruens and

Tabellaria floculossa (Patrick, 1996), with moderate to high oxygen levels as indicated by Pinnularia acrosphaeria and Gomphonema truncatum (van Dam, Mertens, &

Sinkeldam, 1994) this conditions could be confirmed with the D.O. (Behar, 1997) and pH measurements taken in the field (Table 4.3) in which the cattle activity observed in the watershed is a contributor to nitrogen levels in the lake.

C. placentula is present in an important abundance. Both are low profile species which prefer to inhabit on macrophytes (Passy, 2007); this is congruent with the increased amount of this type of organisms in the littoral zone of Lake Siscunsí in the present.

 Delta

The modern water diatom assemblage found in the delta of Lake Siscunsí (Figure 4.6) indicates moderated to high oxygen contents as per suggested by E. biggiba, P. frequentissimum, and Synedra Ulna (van Dam et al., 1994), which coincides with the

D.O measurements taken in the field. In the literature, this diatom association also indicates more alkaline conditions, however the pH measured was close to neutrality

(pH=7.78) (Table 4.4).

 Transitional zones

The diatom assemblage in transitional zones (profundal-littoral) are mainly composed by N. trivialis, G. truncatum and P. pseudoconstruens which according to Passy (2007) are motile species which are highly competitive in nutrient rich environments (as of today in Lake Siscunsí), and can easily adapt to different type of conditions (Danielson,

2009).

 Profundal zone

In this lake zone, a similar diatom assemblage to the transitional zone was observed, however, there was more diversity in diatom species of the motile profile which is congruent with the adaptation to different habitats and luminosity mentioned by Passy

(2007) and Gottschalk and Kahlert (2012)

5.2 Integrated palaeolimnological interpretation

5.2.1 Period 1a ~2,854-1,976 cal yr BP (650-460 cm)

As suggested by high values of MS and sand, erosion was high, and thus an important process at the start of this period; however, the decreasing trend towards ~2,335 cal yr

BP, suggests an environmental transition from high to lower erosive and hydraulic conditions. This can be explained by a contraction of the littoral zone in Lake Siscunsí, this has been observed in several tropical lakes when the finer sediments proportion increases (e.g. Bird et al., 2018; Bogotá-A et al., 2011; Cardozo et al., 2014).

Primary productivity in the lake was the lowest in the whole record (indicated by TOM and Chl 푎), and for the most part it was driven by lacustrine algae and possibly by some

75 terrestrial inputs at the beginning of the period, as indicated by C/N and δ13C values.

C/N, δ15N, total C and total N, followed the same decreasing pattern as the one observed in MS, which indicates that OM at the start of the period originated from allochthonous sources and, around ~ 2,335 cal yr BP, OM was produced mainly within the lake. Lower values of MS towards ~2,335 cal yr BP and stable values after that time, also indicate less inwash of allochthonous material reaching into the lake (Caballero et al., 2002).

Water conditions were fresh to brackish. The abundance of benthic and epiphytic species indicates that water levels were relatively low during most of the period. The water was oligo to mesotrophic as suggested by total N, total C and supported by the abundance of S. construens, S. pinnata, C. placentula, E. adnata, G. affine,

R.abbreviata, E. turgida and P. frequentisimum (see Appendix B). Well aerated waters are indicated by S. construens (Gasse, 1986) and E. adnata according to the modern

DO measurements in Lake Siscunsí. The dominance of high profile diatoms could be associated with low turbidity in the water column, and high luminosity reaching the bottom of the lake (Passy, 2007). The proportion of total carbonates was the highest during the whole record, suggesting possible periods of increased water evaporation

(Rhodes et al., 1996). In summary, this period records a shift from higher energy and runoff conditons to quieter conditions with less influx of allochthonous material getting into the lake (Figure 5.1).

Figure 5.1 Lake Siscunsí dynamics from ~2,854 to 1,976 cal yr BP (Period 1a)

5.2.2 Period 1b ~1,976-1,572 cal yr BP (460-345cm)

Erosion rates were stable as marked by MS, which let accumulation rates to be higher than in the previous period, they remained constant at around 0.6 cm/year. This suggests calm conditions in the water column. This phenomenon allowed heavier material to accumulate (sand), thus the littoral zone expanded, causing a more stratified lake.

The increase in accumulation rates is also related with higher accumulation of total C and total N coming from allochthonous C3 sources. However, autochthonous material

(Chl 푎 and δ13C), indicates reduced algal production, and more N fixation (δ15N). Overall productivity levels in the lake were similar to the previous period.

Grain size, low profile and epiphytic diatoms indicate that water level was higher than in period 1a. Most likely, a developed and expanded littoral zone was inhabited by diatoms belonging to the low profile guild, which thanks to its short stature and its capacity to adhere to the substrate are less impacted by these kind of changes (Gottschalk &

Kahlert, 2012), this is congruent with the abundance of the low profile diatom guild dominating the littoral zone nowadays. The increase in epiphytic diatoms is also indicating an increase of submerged vegetation as indicated by Cocconeis placentula a diatom known for inhabiting in macrophytes (Recasens, Ariztegui, Maidana, &

Zolitschka, 2015).

The dominance of low profile diatoms can be explained by the rising levels of nutrients, an increase in current velocity and generally high disturbance with higher turbidity

(Passy, 2007).These conditions and the increased amount in diatoms that can live in both, benthonic and planktonic habitats support the development of a littoral area, as

78 low profile diatoms are tolerant to high disturbance (Zelnik, Balanč, & Toman, 2018).

The decrease in Staurosira sp nov. could be due to the increase in water levels and/or due to the increase of submerged vegetation; the decrease of this species and the increase in Cocconeis placentula, could also suggest possible competition between these diatom species.

Water conditions were fresh to brackish and similar conditions to the previous period are maintained during this time. The elevated nitrogen concentrations during this period might indicate more mesotrophic conditions caused by active decomposition of organic matter, and the accumulation of allochthonous OM, which is supported by the highest abundance of S. construens, C. placentula, due to their preference for waters with high content of nutrients (Kapetanovic, 2007; Kelly, Penny, & Whitton, 1995).

In summary, during this period the lake was under higher hydraulic conditions, high water levels with increased influx of allochthonous material getting into the lake (Figure

5.2).

Figure 5.2 Lake Siscunsí dynamics from ~1,976 to 1,572 cal yr BP (Period 1b)

5.2.3 Period 2 ~1,572-349 cal yr BP (345-75cm)

MS and accumulation rates were constant from ~1,572 until ~1,035 cal yr BP, however the sand proportion decreased and parallel to that, silt increased, which can be result of a change to lower energy, causing sediment focusing or resuspension (Blais & Kalff,

1995; Dearing, 1997; Lehman, 1975), turning into a shortening of the littoral zone until

~535 cal yr BP.

Chl 푎 showed that primary productivity in lake was the highest of the whole record and

C/N ratio below 10, indicates that during this period, productivity was generated in situ.

According to δ13C, this autochthonous sources, were possibly a combination of CAM-C3 flexible macrophytes and phytoplankton (Boom et al., 2014). This also explains the increase in total N and δ15N, due to the capacity of these macrophytes to fix N (Keeley,

1998). Total Organic Matter remained high and constant for the longest time in the record. The high nitrogen concentrations possibly favoured the growth of S. anceps, S. pupula, P. acrosphaeria, S. acidoclinata (Appendix B). Oxygen levels increased (S. anceps), and there is a high probability of turbid waters (less abundance of P. frequentissimum) during this period (Schneck, Torgan, & Schwarzbold, 2007). All these conditions, along with the presence of S. anceps, S. acidoclinata, S. pupula, P. acrosphaeria, and T. floculosa indicate a meso-eutophic state (Mora, Carmona, &

Cantoral-Uriza, 2015; van Dam et al., 1994), which is supported by the high tolerance to organic pollution that these last two diatom species showed nowadays at Lake Siscunsí.

The return of Staurosira sp. nov could be due to the increased nitrogen concentration and a low abundance in Cocconeis placentula (possible competence).

From the start of this period until ~535 cal yr BP, grain size and the dominance of the diatoms of high profile and benthic species suggest shallow waters due to their capacity to live in the biofilm, demand for high luminosity and an increased amount of macrophytes (Gottschalk & Kahlert, 2012; Passy, 2007).

After ~535 cal yr BP and until the end of this period, accumulation rates decreased, erosion rates increased and sand proportion increased which suggests an increased sedimentological energy and possibly the bypass of sediment. This shift, marks the transition towards period 3, where all variables change: OM starts to decrease slightly, while total N, total C and Chl 푎 decrease abruptly. High profile diatoms decrease and low profile diatoms increase its abundance. All these conditions suggest an increase in water level, a return to more mesotrophic conditions and more inwash coming from the watershed as per indicated by MS, probably as a result of increased precipitation in the watershed.

In summary, during this period the lake presented two different stages, the lowest water levels of the whole record with sediment resuspension, and higher water levels with increased hydraulic energy (Figure 5.3).

Figure 5.3 Lake Siscunsí dynamics from ~1,572 to 349 cal yr BP (Period 2)

5.2.4 Period 3 ~349- present (75-0cm)

MS indicates that erosion occurred continuously during this period; perhaps this is not a natural process but a result of sediment mixing in the lake caused by the construction of a weir in the early 2000s (Redacción el tiempo, 2000) leading to register the lowest accumulation rates in the whole record that could have affected the sediment sorting in the top core samples. However this sediment mixing did not affect the calibration of the diatom stratigraphy for the reconstruction in very shallow lakes (Larsen & MacDonald,

1993), like Lake Siscunsí.

Total organic matter and total carbonates remained relatively stable but lower than in

Period 2 and higher than in 1a. C/N ratio suggest that at the start of this period productivity was generated by allochthonous sources, but as the time passed by, heavier δ13C indicate that CAM macrophytes were not anymore the principal contributors to the productivity of the lake (like in zone 2), and that phytoplankton had the greatest contribution during this period. δ15N indicated ligther signatures, suggesting that the abundance of CAM macrophytes decreased in the lake, thus less N, was fixated into the sediment, this is supported by the lowering trend in total N.

The abundance of the only true planktonic and motile diatoms in the record, suggest mixing water conditions in the lake as well as an increase in the water level.

Grain size suggest that the littoral zone was reduced from ~349 to ~135 cal yr BP and that quieter sedimentological conditions prevailed. After this time until the present they started to increase as indicated by the sand trend. This could reflect sediment mixing during the construction of the weir in recent years, which strengthens the evidence of

84 the increasing trend in erosion seen in MS. The appeareance of F. tenera might be also indicative of turbulent conditions (Corella et al., 2011; Reynolds, 1976), and its association with P. frequentissimum (a periphytic species) supports the mixing of the littoral area (Corella et al., 2011).

These mixing conditions are associated with the motile diatoms dominance at the start of this period due to its resistance to live under that kind of environmental pressure

(Passy, 2007; Svensson, Norberg, & Snoeijs, 2014). Its ease of adaptation could be also associated with a changing environment, which is reflecting variations in water levels and light (Svensson et al., 2014). The modern autecology of these taxa can be totally transferred to the pre-European late Holocene because in this lake the anthropogenic pressure was not high enough to cause an abrupt change in environmental conditions until the construction of the weir.

Productivity was originated mainly from autochthonous sources during this period, and it followed 2 main trends as indicated by Chlorophyll 푎 and total C. The first one from

~349 to ~135 cal yr BP when productivity decreased, and the second one from ~ 135 to present when productivity increased. The decrease in the first one can be explained by the constant disturbance in the lake (O’Beirne et al., 2017), possibly by the construction of the weir while the second trend can reflect an adaptation to the new environment.

During the whole period the water in the lake was fresh-brackish. The oxygen decreased, and the trophic state was meso-eutrophic as indicated by total N, total C and diatoms species that have been found under these trophic levels S. pupula, C.

85 placentula, E. adnata, G. affine, P. frequentisimum, G. truncatum, N. viridula, N. trivialis,

S. phoenicenteron, G. accuminatum (van Dam et al., 1994).

In summary, this period was under high disturbance as indicated by MS, high energy prevailed in a meso-eutrophic environment (Figure 5.4).

Figure 5.4 Lake Siscunsí dynamics from ~349 cal yr BP to present (Period 3)

5.2.5 Inferred palaeoclimatic conditions

According to the proxies analyzed in this record, Lake Siscunsí showed low water levels from ~2,854 to ~1,976 cal yr BP (zone 1a) and from ~1,572 to ~535 (zone 2), while high water levels from ~1,976 to ~1,572 cal yr BP (zone 1b) and from ~535 cal yr BP until present (zone 3). These conditions recorded in the lake, are fingerprinting what happened in the watershed during the last ca. 2,854 years as a result of local and/or regional phenomena. Thus, the intention of the following paragraphs is to explain how climatic teleconections, topography, local and regional factors are affecting the dynamics of Lake Siscunsí and if the anthropogenic factor was an influence on it.

It has been said that South America is one of the regions that is more prone to be impacted by ENSO, causing significant changes in rainfall trends (Staupe-Delgado &

Glantz, 2017), and in this sense, ENSO has been the major factor controlling precipitation anomalies in the Colombian Andes in the last 2,000 years (Flantua et al.,

2016). In this region this phenomenon represents warm and dry conditions during ‘El

Niño’ phase and cold and humid conditions during ‘La Niña’ phase, causing below or above average precipitation and warm air temperature (Staupe-Delgado & Glantz,

2017).

Flantua (2016) stated that ENSO has been controlling precipitation in Colombia during the last 2,000 years (Flantua et al., 2016), however due to the generation of different microclimates origined by topographical settings in the region (e.g. mountain slopes, hills), it is hard to take for granted that this phenomenon is directly affecting Siscunsí. In order to confirm if Lake Siscunsí dynamics has been controlled by ENSO, the

88 precipitation of 2 hydrometeorological stations close to lake Siscunsí (Monguí and

Mongua), was plotted against ‘El Niño’ years (Figures 5.5 and 5.6).

Figures 5.5 and 5.6 show the observed precipitation, as a sum of the trend, the seasonal precipitation and the random values (noise). In which the ‘trend’, represents the precipitation without considering the noise and the seasonal precipitation. These figures let us observe that ENSO has been an important factor controlling the precipitation trend in Lake Siscunsí watershed, causing decreases en precipitation during ‘El Niño’ years at least during the last 60 years, and that the dry island effect in

Lake Siscunsí is not strong enough to show an antiphase of what is occurring in other water bodies located in the lowlands.

Figure 5.5 Precipitation at Monguí and ‘El Niño’ years from 1962 to 1986

Figure 5.6 Precipitation at Mongua and ‘El Niño’ years from 1980 to 1992

In Northern South America, the climatic variability has been associated with the ENSO at interannual scales and with the migration of the ITZC in annual scales, causing lower stream flows and low precipitation rates during El Niño years (Garreaud & Aceituno,

2001; Vuille, Bradley, & Keimig, 2000). These phenomena have been driving the climatic forces in different parts of the eastern Colombian cordillera. For instance in

Lake Fuquene ( 2,540 m asl, Figure 2.1) (Vélez et al., 2003), in Pantano de Vargas

(2,510m asl, Figure 2.1) (Gómez et al., 2007) and in Lake Siscunsí, the dry conditions recorded after ~2,500 cal yr BP correlate with stronger and higher frequency of ENSO activity (Moy, Seltzer, Rodbell, & Anderson, 2002).

Later, during the period between ~1,976 and 1,572 cal yr BP, in lake Siscunsí high water levels were recorded, coinciding with wetter conditions at Paramo de Frontino

(Muñoz et al., 2017) and Turbera de Calostros (Bosman et al., 1994), and opposite to the dry conditions reported by Bird et. al. (2017) at Laguna de Ubaque after ~2100 cal yr BP. Thus, during this time span, Lake Siscunsí seemed to have the same wet period as the other records, except Ubaque, possibly as a result of its location in a hollow within the Eastern cordillera.

In Lake Siscunsí, the lowest lake levels (the driest period), were found between ~1,572 and ~535 cal yr BP, which could be suggesting conditions brought about by ENSO, which showed higher variability during the MCA (from ca. ~1,100 to ~700 yr BP) (Ledru et al., 2013). In this terms, Moy et. al (2002) have reported periods of higher frequency of ENSO having its peak at ~1,200 cal yr BP when ENSO events were more common.

This dry conditions were also recorded since ~1,600 14C yr BP until the present in Lake

Fúquene (Vélez et al., 2003).

After ~535 cal yr BP, until ~349 cal yr BP higher water levels were recorded, also associated with ENSO in its ‘El Niño’ phase, which had a weaker signal during the LIA

(ca. ~400 -150 yr BP) in the tropics (Ledru et al., 2013).

The periods described above, were parallel to the development of the Herrera and Early

Muisca periods, however in Lake Siscunsí erosion rates did not showed large changes that could be related to human impacts, thus it can be inferred that this lake dynamics until ~349 was driven only by natural forces.

From ~349 to the present, Lake Siscunsí showed an increase in water levels, however this increase could not only reflect natural causes but human impact (due to the abrupt change in erosion rates), making the climatic interpretation difficult. In 2000 a weir was built in order to control water level, and as part of tourism in the lake fish were introduced and a stone pathway was also built. There are late Holocene records that range in this time span (e.g Hoover, 2017), however is hard to correlate them to Lake

Siscunsí in this recent period because the water level is totally altered by civil structures and the uppermost part of the lake sediments can contain reworked material originated during the construction of the weir.

5.3 Concluding remarks

Lake Siscunsí reflected four different periods of climatic conditions throughout time, in which dry conditions prevailed from ~2,854 to ~1,976 cal yr BP and from ~ 1,572 to

~535 cal yr BP, and wetter conditions were present from ~1,976 to ~1,572 cal yr BP and from ~535 to ~349 cal yr BP.

Climatic conditions reflect ENSO stronger activity during the MCA and weakened signals during the LIA, which in Lake Siscunsí meant dry and wet conditions respectively, suggesting that this water body was mainly influenced by regional precipitation patterns controlled by ENSO frequency and intensity and global climatic patterns (MCA and LIA).

During the last 2,800 years, primary productivity has been mainly driven by C3, C3/CAM plants and phytoplankton in Lake Siscunsí. N and C were present in proportional balanced amounts, except during period 2 when N increased at a fastest rate when productivity seem to be originated in lake.

Although different civilization groups were inhabiting the highlands of Colombia, the different proxies show that human impact did not have a direct effect on Lake Siscunsí until the early 2,000s. In periods 1b and 2, one would expect that after the domestication and agricultural practices were well established, nitrogen levels would increase, but this is not the case in Lake Siscunsí. MS indicated very stable erosion rates in the lake before ~835 cal yrs BP, suggesting that humans were not directly affecting the lake, until period 3 when the weir was built.

Nowadays the littoral zone of Lake Siscunsí is showed neutral-alkaline pH (7.0-8.2) with good levels of oxygen to be a good habitat for flora and fauna. The modern diatom species found in the lake showed to be resistant to increased organic pollution (inferred from the fish introduction to the lake), and the diatom guilds (high, low, motile) in the lake were clearly good indicators of the different physical conditions of the zones where

94 they inhabit, e.g. marked presence of motile guild close to the weir due to its resistance to live under high current velocities.

Lake Siscunsí must have been an isolated water body given its endemic flora, and different from other studies in the Colombian Eastern Cordillera, it was not impacted by humans until the last ~350 years.

Bibliography

Abbott, M., & Stafford, T. (1996). Radiocarbon Geochemistry of Ancient and Modern

Arctic Lakes, Baffin Island. Quaternary Research, 45, 300–311.

Akaike, H. (1973). Information theory as an extension of the maximum likelihood

principle. In F. Petrov, B.N., Csaki (Ed.), Second International Symposium on

Information Theory (pp. 267–281). Budapest, Hungary: Akademiai Kiado.

Amador, J. A., Alfaro, E. J., Lizano, O. G., & Magaña, V. O. (2006). Atmospheric forcing

of the eastern tropical Pacific: A review. Progress in Oceanography, 69(2–4), 101–

142.

Andrews, J. E., Riding, R., & Dennis, P. F. (1993). Stable isotopic compositions of

Recent freshwater cyanobacterial carbonates from the British Isles: local and

regional environmental controls. Sedimentology, 40(2), 303–314.

Bao, R., Hernández, A., Sáez, A., Giralt, S., Prego, R., Pueyo, J. J., Moreno, A., Valero-

Garcés, B. L. (2015). Climatic and lacustrine morphometric controls of diatom

paleoproductivity in a tropical Andean lake. Quaternary Science Reviews, 129, 96–

110.

Barinova, S. S., Nevo, E., & Bragina, T. M. (2011). Ecological assessment of wetland

ecosystems of northern Kazakhstan on the basis of hydrochemistry and algal

biodiversity. Acta Botanica Croatica, 70(2), 215–244.

Battarbee, R. W. (1986). Diatom analysis. In B. E. Berglund (Ed.), Handbook of

Holocene palaeoecology and palaeohydrology (pp. 527–570). Chinchester: John

Wiley & Sons.

Behar, S., (1997). Testing waters: Chemical and Physical Vital Signs of a River (River

Watch). Montpelier, VT: River Watch Network.

Behling, H., & Hooghiemstra, H. (1999). Environmental history of the Colombian

savannas of the Llanos Orientales since the Last Glacial Maximum from lake

records El Pinal and Carimagua. Journal of Paleolimnology, 21(4), 461-476.

Beltrán-Beltrán, L. C. (2008). La definición de paisajes productivos en la Sabana de

Bogotá. Journal of Cultural Heritage Studies, 21(1), 26–43.

Bird, B. W., Rudloff, O., Escobar, J., Gilhooly, W. P., Correa-Metrio, A., Vélez, M., &

Polissar, P. J. (2018). Paleoclimate support for a persistent dry island effect in the

Colombian Andes during the last 4700 years. The Holocene, 28(2), 217-228.

Blaauw, M., & Christen, J. (2011). Flexible Paleoclimate Age-Depth Models Using an

Autoregressive Gamma Process. Bayesian Analysis, 6(3), 457–474.

Blais, J. M., & Kalff, J. (1995). The influence of lake morphometry on sediment focusing.

Limnology and Oceanography, 40(3), 582–588.

Bogotá-A., R. G., Hooghiemstra, H., & Berrio, J. C. (2016). North Andean environmental

and climatic change at orbital to submillennial time-scales: Vegetation, water-levels

and sedimentary regimes from Lake Fúquene between 284 and 130 ka. Review of

Palaeobotany and Palynology, 226, 91-107.

Bogotá-A, R. G., Groot, M. H. M., Hooghiemstra, H., Lourens, L. J., van der Linden, M.,

& Berrio, J. C. (2011). Rapid climate change from north Andean Lake Fúquene

pollen records driven by obliquity: Implications for a basin-wide biostratigraphic

zonation for the last 284 ka. Quaternary Science Reviews, 30(23–24), 3321–3337.

Boom, A., Carr, A. S., Chase, B. M., Grimes, H. L., & Meadows, M. E. (2014). Leaf wax

n-alkanes and δ13C values of CAM plants from arid southwest Africa. Organic

Geochemistry, 67, 99–102.

Bordovskiy, O. K. (1965). Sources of organic matter in marine basins. Marine Geology,

3(1–2), 5–31.

Bosman, A. F., Hooghiemstra, H., & Cleef, A. M. (1994). Holocene mire development

and climatic change from a high Andean Plantago rigida cushion mire. Holocene,

4(3), 233–243.

Bridges, E. M. (1990). World Geomorphology. Cambridge: Cambridge University Press.

Bruhns, K. (1994). Ancient South America (First). New York: Cambridge University

Press.

Caballero, M., Ortega, B., Valadez, F., Macias, L., Sugiura, Y., & Metcalfe, S. (2002).

Sta. Cruz Atizapán : a 22-ka implications for the late Holocene human occupation in

the Upper Lerma Basin , Central Mexico. Holocene, 186(3–4), 217–235.

Cantonati, M., & Pipp, E. (2000). Longitudinal and seasonal differentiation of epilithic

diatom communities in the uppermost sections of two mountain spring-fed streams.

Internationale Vereinigung Fur Theoretische Und Angewandte Limnologie

Verhandlungen, 3(27), 1591–1595.

Cardozo, A. Y. V., Gomes, D. F., da Silva, E. M., Duque, S. R. E., Rangel, J. O. C.,

Sifeddine, A., Turcq, Bruno, Albuquerque, A. L. S. (2014). Holocene

paleolimnological reconstruction of a high altitude Colombian tropical lake.

Palaeogeography, Palaeoclimatology, Palaeoecology, 415, 127–136.

Chagué-Goff, C., Dawson, S., Goff, J. R., Zachariasen, J., Berryman, K. R., Garnett, D.

L, Waldron, H.M., Mildenhall, D. C. (2002). A tsunami (ca. 6300 years BP) and other

Holocene environmental changes, northern Hawke’s Bay, New Zealand.

Sedimentary Geology, 150(1–2), 89–102.

Clark, M. (2013). Generalized additive models: Getting Started with Additive Models In

R. Center for social Research. University of Notre Dame.

Coe, M., Benson, E., & Snow, D. (1986). Atlas of ancient America. New York: Facts on

File.

Coleman, D. C., & Fry, B. (1991). Carbon Isotope Techniques. Carbon Isotope

Techniques. San Diego, California: Elsevier.

Coletta, P., Pentecost, A., & Spiro, B. (2001). Stable isotopes in charophyte

incrustations: Relationships with climate and water chemistry. Palaeogeography,

Palaeoclimatology, Palaeoecology, 173(1–2), 9–19.

Corella, J. P., Amrani, A. El, Sigró, J., Morellón, M., Rico, E., & Valero-Garcés, B. L.

(2011). Recent evolution of Lake Arreo, northern Spain: Influences of land use

change and climate. Journal of Paleolimnology, 46(3), 469–485.

Correal-Urrego, G. (1990). Evidencias culturales durante el Pleistoceno y Holoceno de

Colombia. Revista de Arqueología Americana, 1, 69–89.

Cremer, H., Gore, D., Hultzsch, N., Melles, M., & Wagner, B. (2004). The diatom flora

and limnology of lakes in the Amery Oasis, East Antarctica. Polar Biology, 27(9),

513–531.

Danielson, T. J. (2009). Protocols for Calculating the Diatom Total Phosphorus Index

(DTPI) and Diatom Total Nitrogen Index (DTNI) for Wadeable Streams and Rivers.

Augusta, ME: Department of Environmental Protection.

Dean, W. E. (1974). Determination of carbonate and organic matter in calcareous

sediments and sedimentary rocks by loss on ignition: Comparison with other

methods. Journal of Sedimentary Petrology, 44, 242–248.

Dearing, J. A. (1997). Sedimentary indicators of lake-level changes in the humid

temperate zone: A critical review. Journal of Paleolimnology, 18(1), 1–14.

Dillehay, T. D. (2000). The settlement of the Americas : a new prehistory. New York:

Basic Books.

Douglas, M. S. V., Smol, J. P., & Blake, W. J. (2000). Summary of paleolimnological

investigations of High Arctic ponds at Cape Herschel, east-central Ellesmere Island,

Nunavut. Bulletin of the Geological Survey of Canada, (529), 257–269.

100

Dueñas, H. (1992). The Paleo ENSO Record in the lower Magdalena Basin, Colombia.

In O. Machare (Ed.), Intern. Symposium Lima (pp. 81–85). Lima: Extended

abstracts.

Fajardo Bernal, S. (2016). Prehispanic and Colonial Settlement Patterns of the

Sogamoso Valley. University of Pittsburgh, Pennsylvania.

Farjalla, V. F., Esteves, F. A., Bozelli, R. L., & Roland, F. (2002). Nutrient limitation of

bacterial production in clear-water Amazonian ecosystems. Hydrobiologia,

489(1990), 197–205.

Faustino, S. B., Fontana, L., Rodrigues Bartozek, E. C., De, C. E., Bicudo, M., De

Campos Bicudo, D., Borges Fontana, F., Luciane, Elaine, C., Rodrigues, B.,

Eduardo, C., Mattos de Campos Bicudo, D. (2016). Composition and distribution of

diatom assemblages from core and surface sediments of a water supply reservoir in

Southeastern Brazil. Biota Neotropica, 16(2), 1–23.

Flantua, S. G., Hooghiemstra, H., Vuille, M., Behling, H., Carson, J. F., Gosling, W.,

Hoyos, I., Ledru, M. P., Montoya, E., Mayle, F., Maldonado, A., Rull, V., Tonello, M.

S., Whitney, B. S., González-Arango,C. (2016). Climate variability and human

impact in South America during the last 2000 years : synthesis and perspectives

from pollen records. Climate of the Past,12, 483-523.

Flügel, E. (2004). Microfacies of Carbonate Rocks: Analysis, Interpretation and

Application. Springer-Verlag Berlin Heidelberg, 976.

101

Fluin, J., Tibby, J., & Gell, P. (2010). The palaeolimnological record from lake

cullulleraine,lower Murray river (South-East Australia): Implications for

understanding riverine histories. Journal of Paleolimnology, 43(2), 309–322.

Fogel, M. L., & Cifuentes, L. A. (1993). Isotope fractionation during primary production.

In: Michael H. Engel, Stephen A. Macko (Eds.), Organic Geochemistry. Topics in

Geobiology, Vol. 11, 73-98, 1993

Gaiser, E. E., & Johansen, J. (2000). Freshwater diatoms from carolina bays and other

isolated wetlands on the atlantic coastal plain of south carolina, U.S.A., with

descriptions of seven taxa new to science. Diatom Research, 15(1), 75–130.

Garcia, Y. C., Martinez, J. I., Velez, M. I., Yokoyama, Y., Battarbee, R. W., & Suter, F.

D. (2011). Palynofacies analysis of the late Holocene San Nicolas terrace of the

Cauca paleolake and paleohydrology of northern South America. Palaeogeography,

Palaeoclimatology, Palaeoecology, 229 (1-2), 298-308.

Garreaud, R. D., & Aceituno, P. (2001). Interannual Rainfall Variability over the South

American Altiplano. Journal of Climate, 14(1987), 2779–2789.

Gasse, F. (1986). East African diatoms. Taxonomy, ecological distribution. In

Bibliotheca Diatomologica (Vol. 11, p. 202 pp.).

Giraldo-Giraldo, M. J., Velásquez-Ruiz, C. A., & Pardo-Trujillo, A. (2017). Late-

Holocene pollen-based paleoenvironmental reconstruction of the El Triunfo wetland,

Los Nevados National Park (Central Cordillera of Colombia). Holocene, 28(2), 183–

194.

102

Gómez, A., Berrío, J. C., Hooghiemstra, H., Becerra, M., & Marchant, R. (2007). A

Holocene pollen record of vegetation change and human impact from Pantano de

Vargas, an intra-Andean basin of Duitama, Colombia. Review of Palaeobotany and

Palynology, 145(1–2), 143–157.

Gómez Tapias, J., Nivia Guevara, Á., Montes Ramírez, N. E., Tejada Avella, M. L.,

Jimémnez Mejía, D. M., Sepúlveda Ospina, M. J., Osorio Naranjo, Gaona Narváez,

J. A., Diedrix, T., Uribe Peña, H., Mora Penagos, M. (2007). Geological Map of

Colombia. Scale 1:1,000,000. Servicio Geologico Colombiano.

González-Carranza, Z., Hooghiemstra, H., & Vélez, M. I. (2012). Major altitudinal shifts

in Andean vegetation on the Amazonian flank show temporary loss of biota in the

Holocene. Holocene, 22(11), 1227–1241.

Gottschalk, S., & Kahlert, M. (2012). Shifts in taxonomical and guild composition of

littoral diatom assemblages along environmental gradients. Hydrobiologia, 649, 41-

56.

Grimm, E. C. (2015). Tilia for windows: pollen spreadsheet and graphics program.

Gruhn, R. (1978). A note on the Excavations at El Bosque, Nicaragua, in 1975. In Early

man in the New World: Occasional papers Department of Anthropology University of

Alberta (pp. 261–262).

Guisan, A., Edwards, T. C., & Hastie, T. (2002). Generalized linear and generalized

additive models in studies of species distributions : setting the scene. Ecological

Modelling, 157, 89–100.

103

Heaton, T. H. E. (1986). Isotopic studies of nitrogen pollution in the hydrosphere and

atmosphere: A review. Chemical Geology: Isotope Geoscience Section.

Heiri, O., Lotter, A., & Lemcke, G. (2001). Loss on Ignition as a Method for Estimating

Organic and Carbonate Content in Sediments: Reproducibility and Comparability of

Results. Journal of Paleolimnology, 25(1), 101–110.

Hermelin, M. (2015). Landscapes and landforms of Colombia. Landscapes and

Landforms of Colombia (First). Cham: Springer.

Herrera, L., Sarmiento, G., Romero, F., Botero, P. J., & Berrio, J. C. (2001). Evolución

Ambiental de la Depresión Momposina (Colombia) desde el Pleistoceno Tardío a los

Paisajes Actuales. Geología Colombiana, 26, 95–121.

Hodell, D. A., Brenner, M., & Curtis, J. H. (2000). Climate change in the Northern

American tropics since the last Age: implications for environment and culture. In

Imperfect Balance: Landscape transformations in the Precolumbian Americas (pp.

13–38). New York: Columbia University Press.

Holper, K. (2018). A ~2500-year high-resolution paleoclimate record from Laguna de

Siscunsi, Boyaca, Colombia. Indiana University-Purdue University Indianapolis.

Hooghiemstra, H., & van der Hammen, T. (1993). Late quaternary vegetation history

and paleoecology of Laguna Pedro Palo (subandean forest belt, Eastern Cordillera,

Colombia). Review of Palaeobotany and Palynology, 77(3–4), 235–262.

104

Hoover, K. M. (2017). The Colombian Holocene and the Identification of the

Anthropocene through a Paleolimnologic Reconstruction of Lake Fúquene in the

Eastern Cordillera of Colombia. University of Regina.

Hostetter, H.P., & Hoshaw, R.W. (1970). Environmental Factors Affecting Resistance to

Desiccation in the Diatom Stauroneis anceps. American Journal of Botany, 57(5),

512–518.

IDEAM. (2016). Atlas Interactivo de Colombia. Retrieved from

http://atlas.ideam.gov.co/presentacion/#

Kammerlander, B., Koinig, K. A., Rott, E., Sommaruga, R., Tartarotti, B., Trattner, F., &

Sonntag, B. (2016). Ciliate community structure and interactions within the

planktonic food web in two alpine lakes of contrasting transparency. Freshwater

Biology, 61(11), 1950–1965.

Kapetanovic, T. (2007). Diatoms of wet habitats in the subalpine belt of Mt. Vranica

(Bosnia and Herzegovina). Proceedings of the 1st Central European Diatom Meeting

2007, 115, 73–78.

Kapetanovic, T., Jahn, R., Redñi, S., & Cari, M. (2011). Diatoms in a poor fen of

Bijambare protected landscape, Bosnia & Herzegovina. Nova Hedwigia, 93(1–2),

125–151.

Keeley, J. E. (1998). CAM photosynthesis in Submerged Aquatic Plants. The Botanical

Review, 64(2), 121–175.

105

Keeley, J. E., & Sandquist, D. R. (1992). Carbon- freshwater plants. Plant, Cell &

Environment, 15(9), 1021–1035.

Kelly, M. G., Penny, C. J., & Whitton, B. A. (1995). Comparative performance of benthic

diatom indices used to assess river water quality. Hydrobiologia, 302(3), 179–188.

Kendall, C. (1998). Tracing Nitrogen Sources and Cycling in Catchments. In C. Kendall

& J. J. McDonnell (Eds.), Isotope Tracers in Catchment Hydrology (pp. 519–576).

Amsterdam: Elsevier.

Kivrak, E., & Uygun, A. (2012). The structure and diversity of the epipelic diatom

community in a heavily polluted stream (the akarçay, Turkey) and their relationship

with environmental variables. Journal of Freshwater Ecology, 27(3), 443–457.

Kuhry, P., Salomons, J. B., Riezebos, P. A., & Van der Hammen, T. (1983).

Paleoecología de los últimos 6.000 años en el area de la Laguna de Otun-El

Bosque. In T. Van der Hammen, A. Pérez, & P. Pinto (Eds.), Studies on Tropical

Andean Ecosystems (1st ed., pp. 227–261). Vaduz: J. Cramer.

Larsen, C.P.S. & MacDonald, G.M. (1993). Lake morphometry, sediment mixing and the

selection of sites for fine resolution palaeoecological studies, Quaternary Science

Reviews, 12(9), 781-792.

Lavallée, D. (2000). The first South Americans- The peopling of a Continent from the

Earliest Evidence to High Culture. Salt Lake City: University of Utah Press.

Leavitt, P. R. (1993). A review of factors that regulate carotenoid and chlorophyll

deposition and fossil pigment abundance. Journal of Paleolimnology, 9(2), 109–127.

106

Ledru, M.-P., Jomelli, V., Samaniego, P., Vuille, M., Hidalgo, S., Herrera, M., & Ceron,

C. (2013). Geoscientific Instrumentation Methods and Data Systems The Medieval

Climate Anomaly and the Little Ice Age in the eastern Ecuadorian Andes. Clim. Past,

9, 307–321.

Lehman, J. T. (1975). Reconstructing the rate of accumulation of lake sediment: The

effect of sediment focusing. Quaternary Research, 5(4), 541–550.

Leng, M. J., Lamb, A. L., Heaton, T. H. E., Marshall, J. D., Wolfe, B. B., Jones, M. D.,

Holmes, J.A. Arrowsmith, C. (2005). Isotopes in lake sediments. In M. J. Leng (Ed.),

Isotopes in palaeoenvironmental research (pp. 147–184). Dordrecht, The

Netherlands: Springer.

Leng, M. J., & Marshall, J. D. (2004). Palaeoclimate interpretation of stable isotope data

from lake sediment archives. In Quaternary Science Reviews (Vol. 23, pp. 811–831).

Li, Y. M., Ferguson, D. K., Zhao, Q., Wang, Y. F., Wang, R. X., & Li, C. Sen. (2015).

Diatom-inferred salinity changes from the Yushe paleolake indicate an aridification

during the Pliocene-Pleistocene transition in north China. Palaeogeography,

Palaeoclimatology, Palaeoecology, 417, 544–553.

Liu, K., & Colinvaux, P. A. (1985). Forest changes in the Amazon Basin during the Last

Glacial Maximum. Nature, 318, 556–557.

Luttge, U. (2010). Ability of crassulacean acid metabolism plants to overcome

interacting stresses in tropical environments. AoB Plants, 1, 1–15.

107

Mackie, E. A. V., Leng, M. J., Lloyd, J. M., & Arrowsmith, C. (2005). Bulk organic δ13C

and C/N ratios as palaeosalinity indicators within a Scottish isolation basin. Journal

of Quaternary Science, 20, 303–312.

Mahapatra, D., Chanakya, H., & Ramachandra, T. (2011). C:N ratio of sediments in a

sewage fed urban lake. International Journal of Geology, 5(3), 86–92.

Maier, M. J. (2014). DirichletReg: Dirichlet Regression for Compositional Data in R.

Research Report Series/Department of Statistics and Mathematics, 125. WU Vienna

University of Economics and Business, Vienna.

Maitland, P. S. (1990). Biology of fresh waters (Second). New York: Chapman and Hall,

Inc.

Marchant, R., Behling, H., Berrio, J. C., Cleef, A., Duivenvoorden, J., Hooghiemstra, H.,

Kuhry, P., Melief, B., Schreve-Brinkman, E., Van Geel, B., Van der Hammen, T.,

Van Reenen, G., Wille, M. (2001). Mid- to Late-Holocene pollen-based biome

reconstructions for Colombia. Quaternary Science Reviews, 20(12), 1289–1308.

Marchant, R., Behling, H., Berrio, J. C., Cleef, A., Duivenvoorden, J., Hooghiemstra, H.,

Kuhry, P., Melief, B., Schreve-Brinkman, E., Van Geel, B., Van der Hammen, T.,

Van Reenen, G., Wille, M. (2002). Pollen-based biome reconstructions for

Colombia at 3000, 6000, 9000, 12000, 15000, and 18000 14C yr ago: Late

quaternary tropical vegetation dynamics. Journal of Quaternary Science, 17(2),

113-129.

108

Marchant, R., Berrio, J. C., Cleef, A., Duivenvoorden, J., Helmens, K., Hooghiemstra,

Kuhry, P., Melief, B., Schreve-Brinkman, E., Van Geel, B.,Van Reenen, G., Van Der

Hammen, T. (2001). A reconstruction of Colombian biomes derived from modern

pollen data along an altitude gradient. In Review of Palaeobotany and Palynology,

117, 79-92.

Marchant, R., & Hooghiemstra, H. (2004). Rapid environmental change in African and

South American tropics around 4000 years before present: A review. Earth-Science

Reviews, 66(3–4), 217–260.

Matthiopoulos, J. (2011). How to be a Quantitative Ecologist: The “A to R” of Green

Mathematics and Statistics. How to be a Quantitative Ecologist: The “A to R” of

Green Mathematics and Statistics. Chichester: John Wiley and Sons.

Mayewski, P. A., Rohling, E. E., Stager, J. C., Karlen, W., Maasch, K. A., Meeker, L. D.,

Meyerson, E., Gasse, F., van Kreveld, S., Holmgren, K., Lee-Thorp, J., Rosqvist, G.,

Rack, F.,Staubwasser, M., Schneider, R., Steig, E. J. (2004). Holocene climate

variability. Quaternary Research, 62(3), 243-255.

McGlynn, G., Mackay, A. W., Rose, N. L., Taylor, R. G., Leng, M. J., & Engstrom, D. R.

(2010). Palaeolimnological evidence of environmental change over the last 400

years in the Rwenzori Mountains of Uganda. Hydrobiologia, 648(1), 109–122.

Meyers, P. A. (1994). Preservation of elemental and isotopic source identification of

sedimentary organic matter. Chemical Geology, 114(3–4), 289–302.

109

Meyers, P. A., & Lallier-Vergès, E. (1999). Lacustrine sedimentary organic matter

records of Late Quaternary paleoclimates. Journal of Paleolimnology, 21(3), 345–

372.

Meyers, P. A., & Teranes, J. L. (2001). Sediment organic matter. In W. M. Last & J. P.

Smol (Eds.), Tracking Environmental Change Using Lake Sediments. Volume 2:

Physical and Geochemical Methods (pp. 239–269). Dordrecht, The Netherlands:

Kluwer Academic Publishers.

Michels, A., Umaña, G., & Raeder, U. (2006). Epilithic diatom assemblages in rivers

draining into Golfo Dulce (Costa Rica) and their relationship to water chemistry,

habitat characteristics and land use. Archiv Für Hydrobiologie, 165(2), 167–190.

Michelutti, N., Blais, J. M., Cumming, B. F., Paterson, A. M., Rühland, K., Wolfe, A. P.,

& Smol, J. P. (2010). Do spectrally inferred determinations of chlorophyll a reflect

trends in lake trophic status? Journal of Paleolimnology, 43(2), 205–217.

Mora, D., Carmona, J., & Cantoral-Uriza, E. A. (2015). Diatomeas epilíticas de la

cuenca alta del río Laja, Guanajuato, México. Revista Mexicana de Biodiversidad,

86(4), 1024–1040.

Morales, E. A. (2005). Observations of the morphology of some known and new

fragilarioid diatoms (Bacillariophyceae) from rivers in the USA. Phycological

Research, 53(2), 113–133.

Morales, E. A., Edlund, M. B., & Spaulding, S. A. (2010). Description and ultrastructure

of araphid diatom species (Bacillariophyceae) morphologically similar to

110

Pseudostaurosira elliptica (Schumann) Edlund et al. Phycological Research, 58(2),

97–107.

Morales, M., Otero, J., van der Hammen, T., Esteban, C., Pedraza C., Rodríguez, N.,

Carol, A., Betancourth, J., Olaya, E., Posada, E., Cárdenas, L. (2007). Atlas de

Páramos de Colombia. Bogotá, D.C: Instituto de Investigación de Recursos

Biológicos Alexander von Humboldt.

Moreira, L. S., Moreira-Turcq, P., Turcq, B., Cordeiro, R. C., Kim, J. H., Caquineau, S.,

Mandeng-Yogo, M., Macario, K., Sinninghe Damsté, J. S. (2013).

Palaeohydrological controls on sedimentary organic matter in an Amazon floodplain

lake, Lake Maracá (Brazil) during the late Holocene. The Holocene, 23 (12), 1903-

1914.

Morris, J. L., le Roux, P. C., Macharia, A. N., Brunelle, A., Hebertson, E. G., & Lundeen,

Z. J. (2013). Organic, elemental, and geochemical contributions to lake sediment

deposits during severe spruce beetle (Dendroctonus rufipennis) disturbances. Forest

Ecology and Management, 289, 78–89.

Moy, C. M., Seltzer, G. O., Rodbell, D. T., & Anderson, D. M. (2002). Variability of El

Niño/Southern Oscillation activity at millennial timescales during the Holocene

epoch. Nature, 420, 162–165.

Münnich, M., & Neelin, J. D. (2005). Seasonal influence of ENSO on the Atlantic ITCZ

and equatorial South America. Geophysical Research Letters, 32(21), 1–4.

111

Muñoz, P., Gorin, G., Parra, N., Velásquez, C., Lemus, D., Monsalve, M. C., & Jojoa, M.

(2017). Holocene climatic variations in the Western Cordillera of Colombia: A

multiproxy high-resolution record unravels the dual influence of ENSO and ITCZ.

Quaternary Science Reviews, 155, 159–178.

Murdock, K. J., Wilkie, K., & Brown, L. L. (2013). Rock magnetic properties, magnetic

susceptibility, and organic geochemistry comparison in core LZ1029-7 Lake

El’gygytgyn, Russia Far East. Climate of the Past, 9(1), 467–479.

Myrbo, A. (2012). Carbon Cycle in Lakes. In R. W. Bengtsson, L., Herschy, R.W.,

Fairbridge (Ed.), Encyclopedia of Lakes and Reservoirs (pp. 121–125). Springer

Science+Business Media.

O’Beirne, M. D., Werne, J. P., Hecky, R. E., Johnson, T. C., Katsev, S., & Reavie, E. D.

(2017). Anthropogenic climate change has altered primary productivity in Lake

Superior. Nature Communications, 8, 3–8.

Oksanen, J., Blanchet, F. G., Friendly, M., Kindt, R., Legendre, P., McGlinn, D.,

Minchin, P.R., O'Hara, R. B., Simpson, G. L., Solymos, P., Stevens, M. H. H.,

Szoecs, E., Wagner, H. (2018). Vegan: Community Ecology Package. Retrieved

from https://cran.r-project.org/package=vegan

Owen, R. B., Renaut, R. W., Hover, V. C., Ashley, G. M., & Muasya, A. M. (2004).

Swamps, springs and diatoms: Wetlands of the semi-arid Bogoria-Baringo Rift,

Kenya. Hydrobiologia, 518(1–3), 59–78.

112

Owen, R. B., Renaut, R. W., & Stamatakis, M. G. (2010). Diatomaceous sedimentation

in late Neogene lacustrine basins of western Macedonia, Greece. Journal of

Paleolimnology, 44(1), 343–359.

Pabón-Caicedo, J. D., Eslava-Ramírez, J. A., & Gómez-Torres, R. E. (2001).

Generalidades de la distribución espacial y temporal de la temperatura del aire y de

la precipitación en Colombia. Bogotá D.C, 4, 47–59.

Passy, S. I. (2007). Diatom ecological guilds display distinct and predictable behavior

along nutrient and disturbance gradients in running waters. Aquatic Botany, 86(2),

171–178.

Patrick, R. (1996). Rivers of the United States, Volume III: The Eastern and

Southeastern States. John Wiley & Sons.

Piperno, D. (1989). Non-AffluentForagers: Resource availability, seasonal shortages,

and the emergence of agriculture in Panamanian tropical forests. In D. Harris & G.

Hillman (Eds.), Foraging and farming: the evolution of plant exploitation (pp. 538–

554). Boston: One World Archaeology.

Polissar, P. J., Abbott, M. B., Wolfe, A. P., Vuille, M., & Bezada, M. (2013).

Synchronous interhemispheric Holocene climate trends in the tropical Andes.

Proceedings of the National Academy of Sciences.

Randsalu-Wendrup, L., Conley, D. J., Carstensen, J., Snowball, I., Jessen, C., & Fritz,

S. C. (2012). Ecological Regime Shifts in Lake Kälksjön, Sweden, in Response to

113

Abrupt Climate Change Around the 8.2 ka Cooling Event. Ecosystems, 15(8), 1336–

1350.

Rau, G. (1978). Carbon-13 depletion in a subalpine lake: Carbon flow implications.

Science, 201(4359), 901–902.

Recasens, C., Ariztegui, D., Maidana, N. I., & Zolitschka, B. (2015). Diatoms as

indicators of hydrological and climatic changes in Laguna Potrok Aike (Patagonia)

since the Late Pleistocene. Palaeogeography, Palaeoclimatology, Palaeoecology,

417, 309–319.

Redacción el tiempo. (2000, January 14). Recuperan Laguna de Siscunsí para

acueducto de las cañas. El Tiempo, p. 1. Retrieved from

http://www.eltiempo.com/archivo/documento/MAM-1297721 on July 2017.

Reynolds, C. S. (1976). Sinking movements of phytoplankton indicated by a simple

trapping methodi. A fragilaria population. British Phycological Journal, 11(3), 279–

291.

Rhodes, T. E., Gasse, F., Ruifen, L., Fontes, J.-C., Keqin, W., Bertrand, P., … Cheng,

Z.-Y. (1996). PALAEO A Late Pleistocene-Holocene lacustrine record from Lake

Manas, Zunggar (northern Xinjiang, western China). Palaeogeography,

Palaeoclimatology, Palaeoecology J.Ch. Fontes Died on February, 120(105), 105–

121.

Robinson, D. (2001). Delta N-15 as an integrator of the nitrogen cycle. Trends in

Ecology & Evolution, 16(3), 153–162.

114

Rodriguez, J. V. (2007). La diversidad poblacional de Colombia en el tiempo y el

espacio: Estudio craneométrico. Revista de La Academia Colombiana de Ciencias

Exactas, Fisicas y Naturales, 31(120), 321–346.

Root, R. B. (1967). The Niche Exploitation Pattern of the Blue-Gray Gnatcatcher.

Ecological Monographs, 37 (4), 317-350.

Roubeix, V., Mazzella, N., Delmas, F., & Coste, M. (2010). In situ evaluation of

herbicide effects on the composition of river periphytic diatom communities in a

region of intensive agriculture. Vie et Milieu, 60(3), 233–241.

Rubel, F., & Kottek, M. (2010). Observed and projected climate shifts 1901-2100

depicted by world maps of the Köppen-Geiger climate classification.

Meteorologische Zeitschrift, 19(2), 135–141.

Saliba, F. M., Ghobara, M. M., Attard, E., & Ellul, B. (2016). Primary Study of the Non-

Marine Epilithic Diatom Communities of Malta and Gozo. International Journal of

Current Microbiology and Applied Sciences, 5(10), 69–78.

Schneck, F., Torgan, L. C., & Schwarzbold, A. (2007). Epilithic diatom community in a

high altitude stream impacted by fish farming in southern Brazil . Study area Epilithic

diatom sampling. Acta Limnologica Brasiliensia, 19(3), 341–355.

Sheibley, R. W., Enache, M., Swarzenski, P. W., Moran, P. W., & Foreman, J. R.

(2014). Nitrogen deposition effects on diatom communities in lakes from three

national parks in Washington state. Water, Air, and Soil Pollution, 225(2), 1-23.

115

Shukla, J. (1984). Predictability of Time Averages: Part II: The Influence of the

Boundary Forcings. In D. M. Burridge & E. Källén (Eds.), Problems and Prospects in

Long and Medium Range Weather Forecasting (pp. 155–206). Berlin, Heidelberg:

Springer Berlin Heidelberg.

Snow, J. W. (1976). The climate of northern South America. In H. E. Landsberg (Ed.),

World survey of climatology (pp. 295–403). Amsterdam: Elsevier.

Sommer, U., Adrian, R., Domis, L. D. S., Elser, J. J., Gaedke, U., Ibelings, B., …

Winder, M. (2012). Beyond the Plankton Ecology Group (PEG) Model: Mechanisms

Driving Plankton Succession. Annual Review of Ecology, Evolution, and

Systematics, 43(1), 429–448.

Spaulding, S. A., de Vijver, B. Van, Hodgson, D. A., McKnight, D. M., Verleyen, E., &

Stanish, L. (2010). Diatoms as indicators of environmental change in antarctic and

subantarctic freshwaters. In The Diatoms: Applications for the Environmental and

Earth Sciences, Second Edition (pp. 267–284).

Staupe-Delgado, R., & Glantz, M. H. (2017). Identifying commonalities between

individual El Niño events. In Bris & Cepin (Eds.), Safety and Reliability -Theory and

Applications- (pp. 1567–1575). London: Taylor & Francis Group.

Stenger-Kovács, C., Lengyel, E., Crossetti, L. O., Üveges, V., & Padisák, J. (2013).

Diatom ecological guilds as indicators of temporally changing stressors and

disturbances in the small Torna-stream, Hungary. Ecological Indicators, 24, 138-147

116

Sun, D., Bloemendal, J., Rea, D. K., Vandenberghe, J., Jiang, F., An, Z., & Su, R.

(2002). Grain-size distribution function of polymodal sediments in hydraulic and

aeolian environments, and numerical partitioning of the sedimentary components.

Sedimentary Geology, 152(3-4), 263-277.

Svensson, F., Norberg, J., & Snoeijs, P. (2014). Diatom cell size, Coloniality and

motility: Trade-Offs between temperature, Salinity and nutrient supply with climate

change. PLoS ONE, 9(10), 1–10.

Talbot, M. R. (2001). Nitrogen isotopes in palaeolimnology. In W. M. Last & J. P. Smol

(Eds.), Tracking environmental change using lake sediments. Volume 2: Physical

and Geochemical Methods (pp. 401–439). Dordrecht, The Netherlands: Kluwer

Academic Publishers.

Thevenon, F., Adatte, T., Spangenberg, J. E., & Anselmetti, F. S. (2012). Elemental

(C/N ratios) and isotopic (δ15Norg, δ13Corg) compositions of sedimentary organic

matter from a high-altitude mountain lake (Meidsee, 2661 m a.s.l., Switzerland):

Implications for Lateglacial and Holocene Alpine landscape evolution. Holocene,

22(10), 1135–1142.

Tranvik, L. J., Downing, J. A., Cotner, J. B., Loiselle, S. A., Striegl, R. G., Ballatore, T.,

Dillon, P., Finlay, K., Fortino, K., Knoll, L.B., Kortelainen, P. L., Kutser, T., Larsen,

S., Laurion, I., Leech, D.M., Leigh McCallister, S., McKnight, D. M., Melack, J.M.,

Overholt, E., Porter, J. A., Prairie, Y., Renwick, W.H., Roland, F., Sherman, B.S.,

Schindler, D.W., Sobek, S.; Tremblay, A., Vanni, M.J., Verschoor, A.M., Von

117

Wachenfeldt, E., Weyhenmeyer, G. A. (2009). Lakes and reservoirs as regulators of

carbon cycling and climate. Limnology and Oceanography, 54(6 part 2), 2298–2314. van Dam, H., Mertens, A., & Sinkeldam, J. (1994). A coded checklist and ecological

indicator values of freshwater diatoms from the Netherlands. Netherlands Journal of

Aquatic Ecology, 28(1), 117–133.

Veblen, T. T., Young, K. R., & Orme, A. R. (2015). The Physical Geography of South

America. Oxford University Press. Retrieved from

https://books.google.ca/books?id=0Q-MY4-nlwwC

Vélez, M., Hooghiemstra, H., Metcalfe, S., Martínez, I., & Mommersteeg, H. (2003).

Pollen-and diatom based environmental history since the Last Glacial Maximum

from the Andean core Fúquene-7, Colombia. Journal of Quaternary Science, 18(1),

17–30.

Vélez, M. I., Berrío, J. C., Hooghiemstra, H., Metcalfe, S., & Marchant, R. (2005).

Palaeoenvironmental changes during the last ca. 8590 calibrated yr (7800

radiocarbon yr) in the dry forest ecosystem of the Patía Valley, Southern Colombian

Andes: A multiproxy approach. Palaeogeography, Palaeoclimatology,

Palaeoecology, 216 (3-4), 279-302.

Velez, M. I., Martínez, J. I., & Suter, F. (2013). Late Holocene history of the floodplain

lakes of the Cauca River, Colombia. Journal of Paleolimnology, 49 (4), 591-604.

118

Vos, P. C., & De Wolf, H. (1994). Palaeoenvironmental research on diatoms in early

and middle Holocene deposits in central North Holland (The Netherlands).

Netherlands Journal of Aquatic Ecology, 28(1), 97–115.

Vuille, M., Bradley, R. S., & Keimig, F. (2000). Climate Variability in the Andes of

Ecuador and Its Relation to Tropical Pacific and Atlantic Sea Surface Temperature

Anomalies. Journal of Climate, 13(Hastenrath 1981), 2520–2535.

Watchorn, M. A., Hamilton, P. B., Anderson, T. W., Roe, H. M., & Patterson, R. T.

(2008). Diatoms and pollen as indicators of water quality and land-use change: A

case study from the Oak Ridges Moraine, Southern Ontario, Canada. Journal of

Paleolimnology, 39(4), 491–509.

Weide, D. M. (2012). Freshwater diatoms as a proxy for late Holocene monsoon

intensity in lac ba bê in the karst region of northern Viêtnam. California State

University, Long Beach.

Werum, M., & Lange-Bertalot, H. (2004). Diatoms in springs from Central Europe and

elsewhere under the influence of hydrologeology and anthropogenic impacts.

Iconographia Diatomologica, 13, 3–417.

Wijmstra, T. A., & Van Der Hammen, T. (1966). Palynological data on the history of

tropical savannas in Northern South America. Leidse Geologische Mededelingen,

38, 71–90.

119

Wood, S. N. (2006). Generalized Additive Models: an introduction with R (First).

Chapman and Hall/CRC. Retrieved from https://reseau-mexico.fr/sites/reseau-

mexico.fr/files/igam.pdf

Zelnik, I., Balanč, T., & Toman, M. J. (2018). Diversity and structure of the

tychoplankton diatom community in the limnocrene spring Zelenci (Slovenia) in

relation to environmental factors. Water (Switzerland), 10(4), 1–12.

Zhang, Z. S., & Mei, Z. P. (1996). Effects of human activities on the ecological changes

of lakes in China. In GeoJournal (Vol. 40, pp. 17–24).

Zong, Y., Lloyd, J. M., Leng, M. J., Yim, W. W.-S., & Huang, G. (2006). Reconstruction

of Holocene monsoon history from Pearl River Estuary, southern China, using

diatoms and carbon isotope ratios. The Holocene, 16, 251–263.

120

APPENDIX A

 R code

 Plots used to validate the statistical analyses

121

R code

################### ######PACKAGES##### ################### install.packages("ggplot2") install.packages("colorspace") install.packages("mgcv") install.packages("gamair") install.packages("vegan") install.packages("analogue") install.packages("devtools") devtools::install_github("gavinsimpson/schoenberg") install.packages("DirichletReg") library(ggplot2) library(mgcv) library(gamair) library(vegan) library(analogue) library(grid) library(gridExtra) library(dplyr) library(tidyr) library('schoenberg') library(DirichletReg) library(splines)

read.csv("LSiscunsi_B.csv",header = TRUE, row.names=2, sep=",")-> lakesiscunsi colnames(lakesiscunsi) lakesiscunsi<- lakesiscunsi[c(1:115),c(3:53,60:70)] head(lakesiscunsi)

# create new dataset without missing data lakesiscunsina<- na.omit(lakesiscunsi) dim(lakesiscunsina) colnames(lakesiscunsina) sppfossil<- lakesiscunsina[c(1:67),c(1:49)] envfossil<-lakesiscunsina[c(1:67), c(50:62)] head(sppfossil) head(envfossil) spp.prop<- tran(sppfossil, method = "proportion") colnames(spp.prop) <- make.cepnames(colnames(spp.prop))

################### 122

######CCA########## ###################

ord <- cca (spp.prop ~ C.N.ratio + Mag.Sus..SI.x.10..5.+ Chl.A..mg.g.+d15N+ d13C,data=envfossil, scale=TRUE) plot(ord, display= c("sp","cn"), xlim=c(-3,3), ylim=c(-1.5,1.5)) points(ord, display="sp", pch=25, cex=0.6, pos=1) summary(ord) anova.cca(ord) anova.cca(ord, by="terms") anova.cca(ord, by="axis")

plot(ord)

colnames1<- make.cepnames(colnames(spp.prop)) head(colnames1) scores(ord)-> ccascores head(ccascores) ####plot sites & env biplot plot(ord, display = c("sites", "bp"))

####plot species & env biplot (the figure looks ok when exported to pdf!) op <- par(mar = c(5,4,1,1)+0.1) plot(ord, type = "p",display = c("species", "bp"), xlim=c(-2.1,3), ylim= c(- 1.7, 1.6)) text(ord, display = "sp", add= TRUE, pos=3, cex=0.6, col="red") par(op) summary(ord)

#################################### ######Generalized Linear Models##### ####################################

###GAM C/N

read.csv("LSiscunsi_B.csv",header = TRUE, sep=",")-> lakesiscunsi2 data.frame(lakesiscunsi2) colnames(lakesiscunsi2) lakesiscunsi.CN<- lakesiscunsi2[c(1:115),c(3,53)] colnames(lakesiscunsi.CN) lakesiscunsi.CN2<- na.omit(lakesiscunsi.CN)

123 lakesiscunsi.CN2$Age<-1950- lakesiscunsi.CN2$Cal.yrs.BP #####Transform to calendar years (AD) dim(lakesiscunsi.CN2)

#####not take values below 5 for a better fit in our model grt5<-subset(lakesiscunsi.CN2,C.N.ratio>5) head(grt5) pltCN<- ggplot(grt5,aes(x= Cal.yrs.BP , y= C.N.ratio))+geom_point()+ labs(x= "Cal yrs BP", y= "C/N ratio") pltCN

#with calendar years#smCN<- gam(C.N.ratio ~ s(Age, k=30, bs = "ad"),data= grt5, family = Gamma(link=log),method = "REML") smCN<- gam(C.N.ratio ~ s(Cal.yrs.BP, k=30, bs = "ad"),data= grt5, family = Gamma(link=log),method = "REML") summary(smCN) layout(matrix(1:4, nrow = 2, byrow = 2)) gam.check(smCN) layout(1) head(smCN) plot(smCN, shade=TRUE, residuals=TRUE, pch=1) plot(fitted(smCN)) ilogit<- family(smCN)$linkinv predCN2<-with(grt5, data.frame(Cal.yrs.BP=seq(min(Cal.yrs.BP), max(Cal.yrs.BP), length = 111))) pred1<-predict(smCN, newdata = predCN2, type = "link",se.fit = TRUE) alpha<-0.05 crit<- qt(1-(alpha/2), df=smCN$df.residual) pred2<-with(pred1, data.frame(fitted= ilogit(fit), upper=ilogit(fit+(crit* se.fit)), lower=ilogit(fit-(crit* se.fit)), Cal.yrs.BP = predCN2$Cal.yrs.BP))

ggplot(grt5, aes(x=Cal.yrs.BP, y= C.N.ratio)) +geom_point() + geom_ribbon(data = pred2, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred2, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) + labs(x = "Cal yrs BP", y = "C/N ratio")

###GAM d13C colnames(lakesiscunsi2) lakesiscunsi.d13C<- lakesiscunsi2[c(1:115),c(3,62)] colnames(lakesiscunsi.d13C)

124 lakesiscunsi.d13C2<- na.omit(lakesiscunsi.d13C) lakesiscunsi.d13C2$Age<-1950- lakesiscunsi.d13C2$Cal.yrs.BP #####Transform to calendar years (AD) dim(lakesiscunsi.d13C2) d13_label<- expression(delta^{13}*C) pltd13C<- ggplot(lakesiscunsi.d13C2, aes(x=Cal.yrs.BP, y= d13C))+ geom_point()+labs (x= "cal yrs BP", y= d13_label) pltd13C

smd13C<- gam(d13C~ s(Cal.yrs.BP, k=30), data= lakesiscunsi.d13C2, method = 'REML') summary(smd13C) layout(matrix(1:4, nrow = 2, byrow = 2)) gam.check(smd13C) layout(1) plot(smd13C, shade=TRUE, residuals=TRUE, pch=1) ilogit1<- family(smd13C)$linkinv family(smd13C) predsmd13C<-with(lakesiscunsi.d13C2, data.frame(Cal.yrs.BP=seq(min(Cal.yrs.BP), max(Cal.yrs.BP), length = 107))) pred3<-predict(smd13C, newdata = predsmd13C, type = "link", se.fit = TRUE) alpha<-0.05 crit<- qt(1-(alpha/2), df=smCN$df.residual) pred4<-with(pred3, data.frame(fitted= ilogit1(fit), upper=ilogit1(fit+(crit* se.fit)), lower=ilogit1(fit-(crit* se.fit)), Cal.yrs.BP = predsmd13C$Cal.yrs.BP)) ggplot(lakesiscunsi.d13C2, aes(x=Cal.yrs.BP, y= d13C)) + geom_point() + geom_ribbon(data = pred4, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred4, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) +labs(x = "Cal yrs BP", y = d13_label)

###GAM d15N colnames(lakesiscunsi2) lakesiscunsi.d15N<- lakesiscunsi2[c(1:115),c(3,61)] colnames(lakesiscunsi.d15N) lakesiscunsi.d15N2<- na.omit(lakesiscunsi.d15N)

125 lakesiscunsi.d15N2$Age<-1950- lakesiscunsi.d15N2$Cal.yrs.BP #####Transform to calendar years dim(lakesiscunsi.d15N2) d15_label<- expression(delta^{15}*N) pltd15N<- ggplot(lakesiscunsi.d15N2, aes(x=Cal.yrs.BP, y= d15N))+ geom_point()+labs (x= "cal yrs BP", y= d15_label) pltd15N smd15N<- gam(d15N~ s(Cal.yrs.BP, k=30), data= lakesiscunsi.d15N2, method = 'REML') summary(smd15N) layout(matrix(1:4, nrow = 2, byrow = 2)) gam.check(smd15N) layout(1) plot(smd15N, shade=TRUE, residuals=TRUE, pch=1) ilogitd15N<- family(smd15N)$linkinv family(smd15N) predsmd15N<-with(lakesiscunsi.d15N2, data.frame(Cal.yrs.BP=seq(min(Cal.yrs.BP), max(Cal.yrs.BP), length = 108))) pred5<-predict(smd15N, newdata = predsmd15N, type = "link", se.fit = TRUE) alpha<-0.05 crit<- qt(1-(alpha/2), df=smd15N$df.residual) pred6<-with(pred5, data.frame(fitted= ilogitd15N(fit), upper=ilogitd15N(fit+(crit* se.fit)), lower=ilogitd15N(fit-(crit* se.fit)), Cal.yrs.BP = predsmd15N$Cal.yrs.BP))

ggplot(lakesiscunsi.d15N2, aes(x=Cal.yrs.BP, y= d15N)) + geom_point() + geom_ribbon(data = pred6, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred6, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE,size = 1) +labs(x = "Cal yrs BP", y = d15_label)

###GAM ChlA colnames(lakesiscunsi2) lakesiscunsi.ChlA<- lakesiscunsi2[c(1:115),c(3,69)] colnames(lakesiscunsi.ChlA) lakesiscunsi.ChlA2<- na.omit(lakesiscunsi.ChlA) lakesiscunsi.ChlA2$Age<-1950- lakesiscunsi.ChlA2$Cal.yrs.BP dim(lakesiscunsi.ChlA2)

126 pltChlA<- ggplot(lakesiscunsi.ChlA2, aes(x=Cal.yrs.BP, y= Chl.A..mg.g.))+ geom_point()+labs (x= "cal yrs BP", y= "Chlorophyll a (mg/g)") pltChlA smChlA<- gam(Chl.A..mg.g.~ s(Cal.yrs.BP, k=30), data= lakesiscunsi.ChlA2, method = 'REML', family = Gamma(link=log)) summary(smChlA) layout(matrix(1:4, nrow = 2, byrow = 2)) gam.check(smChlA) layout(1) plot(smChlA, shade=TRUE, residuals=TRUE, pch=1) ilogitChlA<- family(smChlA)$linkinv family(smChlA) predsChlA<-with(lakesiscunsi.ChlA2, data.frame(Cal.yrs.BP=seq(min(Cal.yrs.BP), max(Cal.yrs.BP), length = 108))) pred7<-predict(smChlA, newdata = predsChlA, type = "link", se.fit = TRUE) alpha<-0.05 crit<- qt(1-(alpha/2), df=smChlA$df.residual) pred8<-with(pred7, data.frame(fitted= ilogitChlA(fit), upper=ilogitChlA(fit+(crit* se.fit)), lower=ilogitChlA(fit-(crit* se.fit)), Cal.yrs.BP = predsChlA$Cal.yrs.BP))

ggplot(lakesiscunsi.ChlA2, aes(x=Cal.yrs.BP, y= Chl.A..mg.g.))+ geom_point() + geom_ribbon(data = pred8, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred8, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) +labs(x = "Cal yrs BP", y = "Chl a (mg/g)")

###GAM TOM colnames(lakesiscunsi2) lakesiscunsi.TOM<- lakesiscunsi2[c(1:115),c(3,63)] colnames(lakesiscunsi.TOM) lakesiscunsi.TOM2<- na.omit(lakesiscunsi.TOM) lakesiscunsi.TOM2$Age<-1950- lakesiscunsi.TOM2$Cal.yrs.BP dim(lakesiscunsi.TOM2) pltTOM<- ggplot(lakesiscunsi.TOM2, aes(x= Cal.yrs.BP, y=X.TOM))+ geom_point()+labs (x= "cal yrs BP", y= " TOM (prop)")

127 pltTOM lakesiscunsi.TOM2$X1TOM<-lakesiscunsi.TOM2$X.TOM/100 smTOM<- gam(X1TOM~ s(Cal.yrs.BP, k=30), data= lakesiscunsi.TOM2, method = 'REML', family = betar()) summary(smTOM) layout(matrix(1:4, nrow = 2, byrow = 2)) gam.check(smTOM) layout(1) plot(smTOM, shade=TRUE, residuals=TRUE, pch=1) print(smTOM) ilogitTOM<- family(smTOM)$linkinv family(smTOM) predsTOM<-with(lakesiscunsi.TOM2, data.frame(Cal.yrs.BP=seq(min(Cal.yrs.BP), max(Cal.yrs.BP),length = 114))) pred11<-predict(smTOM, newdata = predsTOM, type = "link", se.fit = TRUE) alpha<-0.05 crit<- qt(1-(alpha/2), df=smTOM$df.residual) pred12<-with(pred11, data.frame(fitted= ilogitTOM(fit), upper=ilogitTOM(fit+(crit* se.fit)), lower=ilogitTOM(fit-(crit* se.fit)), Cal.yrs.BP = predsTOM$Cal.yrs.BP))

ggplot(lakesiscunsi.TOM2, aes(x=Cal.yrs.BP, y= X1TOM))+ geom_point() + geom_ribbon(data = pred12, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred12, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE,size = 1) +labs(x = "Cal yrs BP", y = "TOM (prop)")

###GAM TC colnames(lakesiscunsi2) lakesiscunsi.TC<- lakesiscunsi2[c(1:115),c(3,64)] colnames(lakesiscunsi.TC) lakesiscunsi.TC2<- na.omit(lakesiscunsi.TC) lakesiscunsi.TC2$Age<-1950- lakesiscunsi.TC2$Cal.yrs.BP dim(lakesiscunsi.TC2) pltTC<- ggplot(lakesiscunsi.TC2, aes(x= Cal.yrs.BP, y=X.TC))+ geom_point()+labs (x= "cal yrs BP", y= "% TC") pltTC lakesiscunsi.TC2$X1TC<-lakesiscunsi.TC2$X.TC/100

128 smTC<- gam(X1TC~ s(Cal.yrs.BP, k=30), data= lakesiscunsi.TC2, method = 'REML', family = betar()) summary(smTC) layout(matrix(1:4, nrow = 2, byrow = 2)) gam.check(smTC) layout(1) plot(smTC, shade=TRUE, residuals=TRUE, pch=1) print(smTC) ilogitTC<- family(smTC)$linkinv family(smTC) predsTC<-with(lakesiscunsi.TC2, data.frame(Cal.yrs.BP=seq(min(Cal.yrs.BP), max(Cal.yrs.BP), length = 114))) pred13<-predict(smTC, newdata = predsTOM, type = "link", se.fit = TRUE) alpha<-0.05 crit<- qt(1-(alpha/2), df=smTC$df.residual) pred14<-with(pred13, data.frame(fitted= ilogitTC(fit), upper=ilogitTC(fit+(crit* se.fit)), lower=ilogitTC(fit-(crit* se.fit)), Cal.yrs.BP = predsTC$Cal.yrs.BP))

ggplot(lakesiscunsi.TC2, aes(x=Cal.yrs.BP, y= X1TC)) +geom_point() + geom_ribbon(data = pred14, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred14, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) +labs(x = "Cal yrs BP", y = "TC (proportion)")

###GAM WT%N colnames(lakesiscunsi2) lakesiscunsi.wtN<- lakesiscunsi2[c(1:115),c(3,70)] colnames(lakesiscunsi.wtN) lakesiscunsi.wtN2<- na.omit(lakesiscunsi.wtN) lakesiscunsi.wtN2$Age<-1950- lakesiscunsi.wtN2$Cal.yrs.BP dim(lakesiscunsi.wtN2) pltwtN<- ggplot(lakesiscunsi.wtN2, aes(x= Cal.yrs.BP, y=wt.N))+ geom_point()+labs (x= "cal yrs BP", y= "wt % N") pltwtN layout(matrix(1:4, nrow = 2, byrow = 2)) grt25<-subset(lakesiscunsi.wtN2,wt.N<2.5) smwtN<- gam(wt.N~ s(Cal.yrs.BP, k=30), data= grt25, method = 'REML', family = Gamma(link=log)) summary(smwtN)

129 gam.check(smwtN) layout(1) plot(smwtN, shade=TRUE, residuals=TRUE, pch=1) print(smwtN) ilogitwtN<- family(smwtN)$linkinv family(smwtN) predswtN<-with(grt25, data.frame(Cal.yrs.BP=seq(min(Cal.yrs.BP), max(Cal.yrs.BP), length = 99))) pred15<-predict(smwtN, newdata = predswtN, type = "link", se.fit = TRUE) alpha<-0.05 crit<- qt(1-(alpha/2), df=smwtN$df.residual) pred16<-with(pred15, data.frame(fitted= ilogitwtN(fit), upper=ilogitwtN(fit+(crit* se.fit)), lower=ilogitwtN(fit-(crit* se.fit)), Cal.yrs.BP = predswtN$Cal.yrs.BP))

ggplot(grt25, aes(x=Cal.yrs.BP, y= wt.N))+ geom_point() + geom_ribbon(data = pred16, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred16, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) +labs(x = "Cal yrs BP", y = "wt % N")

###GAM WT%C colnames(lakesiscunsi2) lakesiscunsi.wtC<- lakesiscunsi2[c(1:115),c(3,71)] colnames(lakesiscunsi.wtC) lakesiscunsi.wtC2<- na.omit(lakesiscunsi.wtC) lakesiscunsi.wtC2$Age<-1950- lakesiscunsi.wtC2$Cal.yrs.BP dim(lakesiscunsi.wtC2) grt205<-subset(lakesiscunsi.wtC2,wt.C<20.5) pltwtC<- ggplot(grt205, aes(x= Cal.yrs.BP, y=wt.C))+ geom_point()+labs (x= "cal yrs BP", y= "wt % C") pltwtC

smwtC<- gam(wt.C~ s(Cal.yrs.BP, k=30), data= grt205, method = 'REML', family = Gamma(link=log)) summary(smwtC) layout(matrix(1:4, nrow = 2, byrow = 2)) gam.check(smwtC) layout(1) plot(smwtC, shade=TRUE, residuals=TRUE, pch=1) print(smwtC)

130

ilogitwtC<- family(smwtC)$linkinv family(smwtC) predswtC<-with(grt205, data.frame(Cal.yrs.BP=seq(min(Cal.yrs.BP), max(Cal.yrs.BP), length = 99))) pred17<-predict(smwtC, newdata = predswtC, type = "link", se.fit = TRUE) alpha<-0.05 crit<- qt(1-(alpha/2), df=smwtC$df.residual) pred18<-with(pred17, data.frame(fitted= ilogitwtC(fit), upper=ilogitwtC(fit+(crit* se.fit)), lower=ilogitwtC(fit-(crit* se.fit)), Cal.yrs.BP = predswtC$Cal.yrs.BP))

ggplot(grt205, aes(x=Cal.yrs.BP, y= wt.C)) + geom_point() + geom_ribbon(data = pred18, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred18, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE,size = 1) +labs(x = "Cal yrs BP", y = "wt % C")

###GAM Magnetic Susc read.csv("LSiscunsi_B.csv",header = TRUE, sep=",")-> lakesiscunsi2 data.frame(lakesiscunsi2) colnames(lakesiscunsi2) lakesiscunsi.MS<- lakesiscunsi2[c(1:115),c(3,54)] colnames(lakesiscunsi.MS) lakesiscunsi.MS2<- na.omit(lakesiscunsi.MS) lakesiscunsi.MS2$Age<-1950- lakesiscunsi.MS2$Cal.yrs.BP dim(lakesiscunsi.MS2)

pltMS<- ggplot(lakesiscunsi.MS2,aes(x= Cal.yrs.BP , y= Mag.Sus..SI.x.10..5.))+ geom_point()+labs (x= "Cal yrs BP", y= "Magnetic Susceptibility (SI)") pltMS layout(matrix(1:4, nrow = 2, byrow = 2)) #with calendar years AD#smCN<- gam(C.N.ratio ~ s(Age, k=30, bs = "ad"),data= grt5, method = "REML") smMS<- gam(Mag.Sus..SI.x.10..5. ~ s(Cal.yrs.BP, k=30),data= lakesiscunsi.MS2,method = "REML") summary(smMS) layout(matrix(1:4, nrow = 2, byrow = 2)) gam.check(smMS)

131 layout(1) head(smMS) plot(smMS, shade=TRUE, residuals=TRUE, pch=1) plot(fitted(smCN))

ilogitMS<- family(smMS)$linkinv predMS2<-with(lakesiscunsi.MS2, data.frame(Cal.yrs.BP=seq(min(Cal.yrs.BP), max(Cal.yrs.BP), length = 112))) pred19<-predict(smMS, newdata = predMS2, type = "link", se.fit = TRUE) alpha<-0.05 crit<- qt(1-(alpha/2), df=smMS$df.residual) pred20<-with(pred19, data.frame(fitted= ilogitMS(fit), upper=ilogitMS(fit+(crit* se.fit)), lower=ilogitMS(fit-(crit* se.fit)), Cal.yrs.BP = predMS2$Cal.yrs.BP))

ggplot(lakesiscunsi.MS2, aes(x=Cal.yrs.BP, y= Mag.Sus..SI.x.10..5.)) + geom_point() + geom_ribbon(data = pred20, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred20, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) +labs(x = "Cal yrs BP", y = "Magnetic Susceptibility (SI)")

#################################### ######Creating groups of graphs##### ####################################

######diatom zones rect1<- data.frame (xmin=-65, xmax=349, ymin=-Inf, ymax=Inf) rect2 <- data.frame(xmin=1572, xmax=1976, ymin=-Inf, ymax=Inf)

###C/N graph lakesiscunsi.CN2 pltCNM<- ggplot()+ geom_point(data=grt5, aes(x=Cal.yrs.BP, y= C.N.ratio))+labs (x=NULL, y= "C/N ratio")+ theme(axis.text.x = element_text(size=10, angle=90), axis.text.y = element_text(size=10, angle=0))+ scale_x_continuous(breaks=seq(-65,2860,200), limits = c(-65,2860))+ scale_y_continuous(limits = c(8.5, 13.5))+ geom_ribbon(data = pred2, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) +

132

geom_path(data = pred2, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) + labs(x = "Cal yrs BP", y = "C/N ratio")+geom_rect(data=rect1,aes(xmin=xmin,xmax=xmax, ymin=ymin,ymax=ymax), alpha=0.1,fill="blue")+ geom_rect(data=rect2,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ theme(axis.text.x=element_blank(),axis.title.x=element_blank(),plot.title=ele ment_blank(), axis.ticks.x=element_blank(),plot.margin = unit(c(0.3,0.5,0.3,0.5), "lines")) pltCNM

###Chl a graph lakesiscunsi.ChlA2 pltChlAM<- ggplot()+ geom_point(data = lakesiscunsi.ChlA2, aes(x=Cal.yrs.BP, y= Chl.A..mg.g.))+ scale_y_continuous(limits=c(0.01,0.08))+labs (x= "Cal yrs BP", y= "Chl a (mg/g)")+ scale_x_continuous(breaks=seq(-65,2860,200), limits = c(-65,2860))+ geom_ribbon(data = pred8, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred8, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) +labs(x = "Cal yrs BP", y = "Chl a (mg/g)")+ theme(axis.text.x = element_text(size=10, angle=90),axis.text.y = element_text(size=10, angle=0))+ geom_rect(data=rect1,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ geom_rect(data=rect2,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ theme(plot.margin = unit(c(0.3,0.5,0.3,0.5), "lines")) pltChlAM

###d15N graph lakesiscunsi.d15N2 pltd15N2M<- ggplot()+ geom_point(data=lakesiscunsi.d15N2, aes(x=Cal.yrs.BP, y= d15N))+

133

scale_x_continuous(breaks=seq(-65,2860,200), limits = c(-65,2860))+ scale_y_continuous(limits=c(0.5,3))+ geom_ribbon(data = pred6, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred6, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) +labs(x = "Cal yrs BP", y = d15_label)+ labs (x= NULL, y= d15_label)+ theme(axis.text.x = element_text(size=10, angle=90), axis.text.y = element_text(size=10, angle=0))+ geom_rect(data=rect1,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ geom_rect(data=rect2,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ theme(axis.text.x=element_blank(),axis.title.x=element_blank(),plot.title=ele ment_blank(), axis.ticks.x=element_blank(),plot.margin = unit(c(0.3,0.5,0.3,0.5), "lines")) pltd15N2M

########d13C graph lakesiscunsi.d13C2 head(lakesiscunsi.d13C2) pltd13C2M<- ggplot()+ geom_point(data=lakesiscunsi.d13C2, aes(x=Cal.yrs.BP, y= d13C))+ scale_x_continuous(breaks=seq(-65,2860,200), limits = c(-65,2860))+ scale_y_continuous(limits = c(-29,-21))+ geom_ribbon(data = pred4, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred4, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) +labs(x = "Cal yrs BP", y = d13_label)+ labs (x= NULL, y= d13_label)+ theme(axis.text.x = element_text(size=10, angle=90), axis.text.y = element_text(size=10, angle=0))+ geom_rect(data=rect1,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ geom_rect(data=rect2,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ theme(axis.text.x=element_blank(),axis.title.x=element_blank(),plot.title=ele ment_blank(), axis.ticks.x=element_blank(),plot.margin = unit(c(0.3,0.5,0.3,0.5), "lines")) pltd13C2M

134

#####TOM graph lakesiscunsi.TOM2 pltTOMM<- ggplot()+ geom_point(data=lakesiscunsi.TOM2, aes(x=Cal.yrs.BP, y= X1TOM))+ scale_x_continuous(breaks=seq(-65,2860,200), limits = c(-65,2860))+ scale_y_continuous(limits = c(0.20, 0.45))+ geom_ribbon(data = pred12, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred12, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) +labs(x = "Cal yrs BP", y = "TOM %")+ labs (x=NULL, y= "TOM %")+theme(axis.text.x = element_text(size=10, angle=90), axis.text.y = element_text(size=10,angle=0))+ geom_rect(data=rect1,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ geom_rect(data=rect2,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ theme(axis.text.x=element_blank(),axis.title.x=element_blank(),plot.title=ele ment_blank(), axis.ticks.x=element_blank(),plot.margin = unit(c(0.3,0.5,0.3,0.5), "lines")) pltTOMM

#####TC graph lakesiscunsi.TC2 pltTCM<- ggplot()+ geom_point(data=lakesiscunsi.TC2, aes(x=Cal.yrs.BP, y= X1TC))+ scale_x_continuous(breaks=seq(-65,2860,200), limits = c(-65,2860))+ scale_y_continuous(limits = c(0, 0.07))+ geom_ribbon(data = pred14, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred14, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) +labs(x = "Cal yrs BP", y = "T Carbonates %")+labs (x=NULL, y= "T Carbonates %")+ theme(axis.text.x = element_text(size=10, angle=90), axis.text.y = element_text(size=10, angle=0))+ geom_rect(data=rect1,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ geom_rect(data=rect2,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+

135

theme(axis.text.x=element_blank(),axis.title.x=element_blank(),plot.title=ele ment_blank(), axis.ticks.x=element_blank(),plot.margin = unit(c(0.3,0.5,0.3,0.5), "lines")) pltTCM

#########wt N lakesiscunsi.wtN2 pltwtNM<- ggplot()+ geom_point(data=grt25, aes(x=Cal.yrs.BP, y= wt.N))+labs (x=NULL, y= "TN (wt %)")+ theme(axis.text.x = element_text(size=10, angle=90), axis.text.y = element_text(size=10, angle=0))+ scale_x_continuous(breaks=seq(-65,2860,200), limits = c(-65,2860))+ scale_y_continuous(limits = c(1.1, 2.2))+ geom_ribbon(data = pred16, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred16, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) +labs(x = "Cal yrs BP", y = "TN(wt %)")+ geom_rect(data=rect1,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ geom_rect(data=rect2,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ theme(axis.text.x=element_blank(),axis.title.x=element_blank(), plot.title=element_blank(), axis.ticks.x=element_blank(),plot.margin = unit(c(0.3,0.5,0.3,0.5), "lines")) pltwtNM

#####wt C plot lakesiscunsi.wtC2 pltwtCM<- ggplot()+ geom_point(data=grt205, aes(x=Cal.yrs.BP, y= wt.C))+labs (x=NULL, y= "TC (wt %)")+ theme(axis.text.x = element_text(size=10, angle=90), axis.text.y = element_text(size=10, angle=0))+ scale_x_continuous(breaks=seq(-65,2860,200), limits = c(-65,2860))+ scale_y_continuous(limits = c(11.8, 20))+ geom_ribbon(data = pred18, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred18, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) +labs(x = "Cal yrs BP", y = "TC (wt %)")+ geom_rect(data=rect1,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+

136

geom_rect(data=rect2,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ theme(axis.text.x=element_blank(),axis.title.x=element_blank(),axis.ticks.x=e lement_blank(), plot.margin = unit(c(0.3,0.5,0.3,0.5), "lines")) pltwtCM

#####MS plot lakesiscunsi.MS2 pltMS<- ggplot()+ geom_point(data = lakesiscunsi.MS2, aes(x=Cal.yrs.BP, y=Mag.Sus..SI.x.10..5.))+ scale_y_continuous(limits = c(-1.57, 0))+labs (x= "Cal yrs BP", y="Magnetic Susceptibility (SI)")+ scale_x_continuous(breaks=seq(-65,2860,200), limits = c(-65,2860))+ geom_ribbon(data = pred20, mapping = aes(x =Cal.yrs.BP, ymin = lower, ymax = upper), fill = "grey", colour =NA, alpha = 0.5, inherit.aes = FALSE) + geom_path(data = pred20, mapping = aes(x = Cal.yrs.BP, y = fitted), inherit.aes = FALSE, size = 1) +labs(x = "Cal yrs BP", y = "Magnetic Susceptibility (SI)")+ theme(axis.text.x = element_text(size=10, angle=90),axis.text.y = element_text(size=10, angle=0))+ geom_rect(data=rect1,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ geom_rect(data=rect2,aes(xmin=xmin,xmax=xmax,ymin=ymin,ymax=ymax),alpha=0.1,f ill="blue")+ theme(plot.margin = unit(c(0.3,0.5,0.3,0.5), "lines")) pltMS

##### plot group 1 ggpl1<- ggplot_gtable(ggplot_build(pltChlAM)) ggpl2<- ggplot_gtable(ggplot_build(pltd13C2M)) ggpl3<- ggplot_gtable(ggplot_build(pltd15N2M)) ggpl4<-ggplot_gtable(ggplot_build(pltTOMM)) ggpl5<-ggplot_gtable(ggplot_build(pltTCM)) maxWidth = unit.pmax(ggpl1$widths[2:3], ggpl2$widths[2:3],ggpl3$widths[2:3], ggpl4$widths[2:3],ggpl5$widths[2:3]) ggpl1$widths[2:3] <- maxWidth ggpl2$widths[2:3] <- maxWidth ggpl3$widths[2:3] <- maxWidth ggpl4$widths[2:3] <- maxWidth ggpl5$widths[2:3] <- maxWidth grid.arrange(ggpl5,ggpl4,ggpl3,ggpl2, ggpl1, ncol=1)

137

##### graph 2 ggpl7<- ggplot_gtable(ggplot_build(pltSandM)) ggpl8<- ggplot_gtable(ggplot_build(pltSiltM)) ggpl9<- ggplot_gtable(ggplot_build(pltClayM)) ggpl10<- ggplot_gtable(ggplot_build(pltPM)) maxWidth = unit.pmax(ggpl7$widths[2:3], ggpl8$widths[2:3], ggpl9$widths[2:3],ggpl10$widths[2:3]) ggpl7$widths[2:3] <- maxWidth ggpl8$widths[2:3] <- maxWidth ggpl9$widths[2:3] <- maxWidth ggpl10$widths[2:3] <- maxWidth grid.arrange(ggpl10,ggpl9,ggpl8,ggpl7, ncol=1)

##### graph 3 ggpl11<- ggplot_gtable(ggplot_build(pltMS)) ggpl12<- ggplot_gtable(ggplot_build(pltCNM)) ggpl13<- ggplot_gtable(ggplot_build(pltwtNM)) ggpl14<- ggplot_gtable(ggplot_build(pltwtCM))

maxWidth = unit.pmax(ggpl11$widths[2:3], ggpl12$widths[2:3], ggpl13$widths[2:3],ggpl14$widths[2:3]) ggpl11$widths[2:3] <- maxWidth ggpl12$widths[2:3] <- maxWidth ggpl13$widths[2:3] <- maxWidth ggpl14$widths[2:3] <- maxWidth grid.arrange(ggpl12,ggpl13,ggpl14,ggpl11, ncol=1

#################################### ######to use for SCHOENBERG######### ####################################

#####C/N fdCN<- fderiv(smCN) head(fdCN) fdCN.ci <- confint(fdCN) head(fdCN.ci)

#################################### ######to use for DIRICHLET ####### ####################################

#### DATABASE WITH AGE and depth lakesiscunsi.dra<- lakesiscunsi2[c(1:115),c(2:3, 66:68)] head(lakesiscunsi.dra)

138

## calendar years (negated) lakesiscunsi.dra <- transform(lakesiscunsi.dra, negCalAge = - Cal.yrs.BP) lakesiscunsi.dra$Clay<-lakesiscunsi.dra$X.Clay/100 lakesiscunsi.dra$Silt<-lakesiscunsi.dra$X.Silt/100 lakesiscunsi.dra$Sand<-lakesiscunsi.dra$X.Sand/100 colnames(lakesiscunsi.dra) rownames(lakesiscunsi.dra) lakesiscunsi.dra<- lakesiscunsi.dra[c(1:115),c(9:6)] head(lakesiscunsi.dra)

DR<- DR_data(lakesiscunsi.dra[,1:3]) layout(1) plot(DR, cex=0.5, a2d=list(c.grid=FALSE)) summary(DR)

plot(rep(lakesiscunsi.dra$negCalAge,3), as.numeric(DR), ylim= 0:1, pch=21,bg=rep(c("deeppink3", "darkturquoise", "lightslategrey"), each=114),xlab="cal yrs BP",ylab="Proportion")

######database for AIC with df####### lake0 <- DirichReg(DR ~ negCalAge , lakesiscunsi.dra) summary(lake0)

##too low####lake1 <- DirichReg(DR ~ bs(negCalAge , df=1), lakesiscunsi.dra) ######summary(lake1)

##too low####lake2 <- DirichReg(DR ~ bs(negCalAge ,df=2), lakesiscunsi.dra) ####summary(lake2) lake3 <- DirichReg(DR ~ bs(negCalAge ,df=3), lakesiscunsi.dra) summary(lake3) lake4 <- DirichReg(DR ~ bs(negCalAge ,df=4), lakesiscunsi.dra) summary(lake4) lake5 <- DirichReg(DR ~ bs(negCalAge ,df=5), lakesiscunsi.dra) summary(lake5) lake6 <- DirichReg(DR ~ bs(negCalAge ,df=6), lakesiscunsi.dra) summary(lake6) lake7 <- DirichReg(DR ~ bs(negCalAge ,df=7),lakesiscunsi.dra) summary(lake7) lake8 <- DirichReg(DR ~ bs(negCalAge ,df=8), lakesiscunsi.dra) summary(lake8) lake9 <- DirichReg(DR ~ bs(negCalAge ,df=9), lakesiscunsi.dra) summary(lake9)

139 lake10 <- DirichReg(DR ~ bs(negCalAge , df=10), lakesiscunsi.dra) summary(lake10) lake11 <- DirichReg(DR ~ bs(negCalAge , df=11), lakesiscunsi.dra) summary(lake11) lake12 <- DirichReg(DR ~ bs(negCalAge ,df=12), lakesiscunsi.dra) summary(lake12) lake13 <- DirichReg(DR ~ bs(negCalAge ,df=13), lakesiscunsi.dra) summary(lake13) lake14 <- DirichReg(DR ~ bs(negCalAge ,df=14), lakesiscunsi.dra) summary(lake14) lake15 <- DirichReg(DR ~ bs(negCalAge ,df=15), lakesiscunsi.dra) summary(lake15) lake16 <- DirichReg(DR ~ bs(negCalAge ,df=16), lakesiscunsi.dra) summary(lake16) lake17 <- DirichReg(DR ~ bs(negCalAge ,df=17), lakesiscunsi.dra) summary(lake17) lake18 <- DirichReg(DR ~ bs(negCalAge ,df=18), lakesiscunsi.dra) summary(lake18) lake19 <- DirichReg(DR ~ bs(negCalAge ,df=19), lakesiscunsi.dra) summary(lake19) lake20 <- DirichReg(DR ~ bs(negCalAge ,df=20), lakesiscunsi.dra) summary(lake20) lake21 <- DirichReg(DR ~ bs(negCalAge ,df=21), lakesiscunsi.dra) summary(lake21) lake22 <- DirichReg(DR ~ bs(negCalAge ,df=22), lakesiscunsi.dra) summary(lake22)

####after 22 It didn’t converge dfreedom<-c(3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22) aicvl<- c(-334.7,-340.7,-336.6,-335.6,-368.1,-358.1,-376.1,-382.1, -386.6,-374.2,-376.9,-372.4,-381.5,-376.6,-400.2,-406.5, -409,-403.1,-415.8,-429) aicdata<-data.frame(dfreedom,aicvl) pltaic<-plot(aicdata)

#########plot sand, silt and clay par(mar = c(4, 4, 4, 4) + 0.1) plot(rep(lakesiscunsi.dra$negCalAge, 3), as.numeric(DR), pch = 21,cex=0.7, bg=rep(c("deeppink3","darkturquoise", "lightslategrey"), each = 115), xlab = "cal yrs BP", ylab = "Proportion",ylim = 0:1,

140

main = "Sediment Composition in Lake Siscunsí")

Xnew <- data.frame(negCalAge = seq(min(lakesiscunsi.dra$negCalAge), max(lakesiscunsi.dra$negCalAge),length.out = 100)) for (i in 1:3) lines(cbind(Xnew, predict(lake22, Xnew)[, i]),col = c("deeppink3",

"darkturquoise",

"lightslategrey")[i],lwd = 2) legend("topright", legend =c("Sand", "Silt","Clay"),lwd = 1, col = c("deeppink3","darkturquoise", "lightslategrey"), cex= 0.6, pt.bg = c("deeppink3", "darkturquoise", "lightslategrey"), pch = 21,bty = "n") par(new = TRUE)

141

Plots used to validate the statistical analyses

 C/N

Normal QQ plot (almost a straight line) (upper left), the distributional assumption is reasonable. Residuals vs linear predictors (upper right), histogram of residuals (lower left). Response vs. fitted values (lower right).

142

 Total carbon

143

 Total nitrogen

144

 δ13C

145

 δ15N

146

 Total Organic Carbon (TOM)

147

 Total carbonates (Tcarb)

148

 Magnetic Susceptibility

149

 Chlorophyll a

150

APPENDIX B

 Autecology of the most representative taxa

 Table B-1 Ecological parameters based on diatom species and percentages

151

Autecology of the most representative taxa

 Cocconeis placentula (Ehrenberg, 1838)

According to Cantonati & Pipp (2000), Cocconeis placentula prefers to inhabit in downstream flows of rivers. It requires light in the water column and it has been found as a tychoplaktonic (Bao et al., 2015), benthic (Bao et al., 2015; Kelly et al., 1995), epilithic (Saliba, Ghobara, Attard, & Ellul, 2016) and epiphytic (Recasens et al., 2015) taxon.

In terms of salinity, it tolerates high levels of salinity and as a consequence it can develop in waters with high conductivity (Saliba et al., 2016). It is tolerant to mild organic pollution and it is commonly found in nutrient rich waters (Kelly et al., 1995). It is often found in brackish to freshwater habitats, however when accompanied by epithemioid species, tends to have a freshwater preference (Chagué-Goff et al., 2002).

 Fragilaria tenera (W. Smith) Lange-Bertalot 1980

Reynolds (1976) observed that this taxon needs turbulent conditions to float in the water column, similar with the high water turbidity requirements reported by Corella et al.

(2011), hence, poor light requirements can be inferred. Fragilaria tenera has been identified as tychoplanktonic (Corella et al., 2011), whereas Kammerlander et al.

(2016), report it is a planktonic species. Although it prefers high salinity levels, it can tolerate low salinity (Moreira et al., 2013). It is also indicative of high nutrient levels in high mountain lakes (Das et al., 2005). Van Dam, Mertens, & Sinkeldam (1994) record it in freshwater habitats with very small concentrations of nitrogen and acidic conditions.

152

When associated with periphytic species, it can indicate reworked material from littoral areas (Corella et al., 2011). Sheibley, Enache, Swarzenski, Moran, & Foreman (2014), reported increases in F. tenera associated with rises in N and P, explained by fish introductions.

 Planothidium frequentissimum (Lange-Bertalot) Lange-Bertalot 1999

Planothidium frequentissimum is identified as periphythic (Randsalu-Wendrup et al.,

2012). Michels, Umaña, & Raeder (2006) propose that this taxon has a better development at low light intensities. It tolerates low salinity levels, therefore it is normally found in fresh-brackish waters (van Dam et al., 1994). It inhabits places with low availability of nutrients with small quantities of total solids and low turbidity (Schneck et al., 2007), however it has been used as an indicator of eutrophication (Rimet et al.

2004; Roubeix, Mazzella, Delmas, & Coste, 2010) due to its tolerance to organic pollution and to its alkalophilous character (van Dam et al., 1994).

 Punctastriata mimetica (Morales, 2005)

This taxon has been reported as meso-eutrophic (Cremer & Koolmees, 2010), however little is known about any other ecological preferences (Weide, 2012). The most recent publications on this species have been focused on its description rather than in any ecological parameter (E. A. Morales, 2005; E. A. Morales, Edlund, & Spaulding, 2010).

 Sellaphora pupula (Kützing) Mereschkovsky 1902

This taxon is known for being benthic, oligohalobous-halophilous, and for tolerating acid waters (Barinova, Nevo, & Bragina, 2011; Watchorn, Hamilton, Anderson, Roe, &

153

Patterson, 2008) although Kivrak & Uygun (2012) found it in neutral waters. Sellaphora pupula in tropical regions is present in both dry and wet seasons even though its dominance is marked during the wet season, it is related to eutrophic-hypertrophic environments. It is highly tolerant to heavy pollution (Mora et al., 2015).

 Stauroneis acidoclinata (Lange-Bertalot & Werum, 2004)

Kapetanovi, Jahn, Redñi, & Cari (2011) found the greatest dominance of this species in a fen, then occurring in waters with low electrolyte levels. It is a benthic diatom (Werum

& Lange-Bertalot, 2004), associated with low conductivity and acidic waters (Faustino et al., 2016).

 Stauroneis anceps (Ehrenberg, 1843)

It is a benthic diatom (Cremer, Gore, Hultzsch, Melles, & Wagner, 2004; Gaiser &

Johansen, 2000) found in swamps, wet meadows, blister mounds (Owen, Renaut,

Hover, Ashley, & Muasya, 2004) and freshwater bodies and soils (Krammer & Lange-

Bertalot,1986) . It can tolerate elevated concentrations of organically bound nitrogen and it normally grows in fresh-brackish waters with a circumneutral pH (van Dam et al.,

1994). It has the ability to survive severe desiccation (Hostetter & Hoshaw, 1970).

 Staurosira construens (Ehrenberg, 1843)

It is known to inhabit littoral areas as an epiphytic taxon (McGlynn et al., 2010; Owen,

Renaut, & Stamatakis, 2010), it has also been reported as a benthic as well as a planktonic species (Spaulding et al., 2010), and as a tychoplaktonic species (McGlynn et al., 2010). It is commonly found in well aerated waters (Gasse, 1986) and in

154 minerotrophic habitats with acidic pH and low conductivity (Kapetanovic, 2007), unlike

Owen et al. (2010), Li et al. (2015) and Vos & De Wolf (1994) who found it in alkaline and neutral waters, and Fluin, Tibby, & Gell (2010) who used it to explain increases in conductivity levels.

155

Table B-1. Ecological parameters based on diatom species and percentages. Adapted from van Dam et al. (1994)

ZONE 1A

10- F. subsalina ALKALOPHILOUS 90% 5- S. construens ALKALOPHILOUS 50% 0- S. pinnata, C. ALKALOPHILOUS 20% placentula pH 0- Others: E. ALKALOPHILOUS 10% adnata,G. affine, R.abbreviata, E. turgida, P. frequentisimum

10- F. subsalina BRACKISH-FRESH 90% 5- S. construens BICARBONATE WATERS 50% 0- S. pinnata, C. FRESH-BRACKISH 20% placentula SALINITY 0- Others: E. FRESH-BRACKISH 10% adnata ,G. affine, R. abbreviata, E. turgida, P. frequentisimum

10- F. subsalina TOLLERATE VERY SMALL CONCENTRATIONS OF ORGANICALLY 90% BOUND NITROGEN 5- S. construens TOLLERATE ELEVATED CONCENTRATIONS OF ORGANICALLY 50% BOUND NITROGEN 0- S. pinnata, C. TOLLERATE ELEVATED CONCENTRATIONS OF ORGANICALLY 20% placentula BOUND NITROGEN NITROGEN 0- Others: E. TOLLERATE SMALL TO ELEVATED CONCENTRATIONS OF 10% adnata,G. ORGANICALLY BOUND NITROGEN affine, R.abbreviata, E. turgida, P. frequentisimum

10- F. subsalina FAIRLY HIGH (ABOVE 75% SATURATION) 90% OXYGEN 5- S. construens CONTINUOUSLY HIGH (ABOUT 100% SATURATION) 50% 0- S. pinnata, C. CONTINUOUSLY HIGH (ABOUT 100% SATURATION)

156

20% placentula 0- Others: E. CONTINUOUSLY HIGH TO MODERATE (ABOVE 50% SATURATION) 10% adnata,G. affine, R.abbreviata, E. turgida, P. frequentisimum

10- F. subsalina α,β - MESOSAPROBOUS 90% 5- S. construens β - MESOSAPROBOUS 50% 0- S. pinnata, C. β - MESOSAPROBOUS 20% placentula SAPROBITY 0- Others: E. α,β - MESOSAPROBOUS 10% adnata,G. affine, R.abbreviata, E. turgida, P. frequentisimum

10- F. subsalina ND 90% 5- S. construens MESO-EUTRAPHENTIC 50% 0- S. pinnata, C. OLIGO-EUTRAPHENTIC TROPHIC 20% placentula STATE 0- Others: E. MESO-EUTRAPHENTIC - MESOTRAPHENTIC 10% adnata,G. affine, R.abbreviata, E. turgida, P. frequentisimum

10- F. subsalina OCCURRING IN WATER BODIES 90% 5- S. construens NEVER OR ONLY RARELY OCCURRING OUTSIDE WATER BODIES 50% 0- S. pinnata, C. MAINLY OCCURRING IN WATER BODIES, ALSO RATHER REGULARLY 20% placentula ON WET AND MOIST PLACES MOISTURE 0- Others: E. MAINLY OCCURRING IN WATER BODIES, ALSO RATHER REGULARLY 10% adnata,G. ON WET AND MOIST PLACES affine, R.abbreviata, E. turgida, P. frequentisimum

157

ZONE 1B

1- F. subsalina ALKALOPHILOUS 80%

20- S. ALKALOPHILOUS 45% construens, C. placentula

0- E. turgida, S. CIRCUMNEUTRAL INDICATED BY S. PUPULA 10% pupula, G. pH affine, R. abbreviata, E. adnata, S. pinnata

0- S. ALKALOPHILOUS 5% acidoclinatta, D. smiithi v. Dilatata

1- F. subsalina BRACKISH-FRESH 80%

20- S. BICARBONATE WATERS, ALKALOPHILOUS 45% construens, C. placentula

0- E. turgida, S. 10% pupula, G.

SALINITY affine, R abbreviata, E. adnata, S. pinnata FRESH-BRACKISH

0- S. FRESH WATERS 5% acidoclinatta, D. smiithi v. Dilatata

158

1- F. subsalina TOLLERATE VERY SMALL CONCENTRATIONS OF 80% ORGANICALLY BOUND NITROGEN

20- S. TOLLERATE ELEVATED CONCENTRATIONS OF ORGANICALLY 45% construens, BOUND NITROGEN C. placentula

0- E. turgida, S. TOLLERATE SMALL TO ELEVATED CONCENTRATIONS OF NITROGEN 10% pupula, G. ORGANICALLY BOUND NITROGEN affine, R. abbreviata, E. adnata, S. pinnata

0- S. ND 5% acidoclinata, D. smiithi v. Dilatata

1- F. subsalina FAIRLY HIGH (ABOVE 75% SATURATION) 80%

20- S. -, MODERATE OXYGEN 45% construens, C. placentula

0- E. turgida, S. FAIRLY HIGH – CONTINUOUSLY HIGH OXYGEN 10% pupula, G. OXYGEN affine, R abbreviata, E. adnata, S. pinnata

0- S. ND 5% acidoclinatta, D. smiithi v. Dilatata

SAPROBITY 1- F. subsalina α,β - MESOSAPROBOUS

159

80%

20- S. β - MESOSAPROBOUS 45% construens, C. placentula

0- E. turgida, S. α,β - MESOSAPROBOUS 10% pupula, G. affine, R abbreviata, E. adnata, S. pinnata

0- S. nd 5% acidoclinata, D. smiithi v. Dilatata

1- F. subsalina ND 80%

20- S. EUTRAPHENTIC 45% construens, C. placentula

0- E. turgida, S. Meso-eutraphentic, mesotraphentic, eutraphentic, oligo- TROPHIC 10% pupula, G. eutraphentic STATE affine, R abbreviata, E. adnata, S. pinnata

0- S. nd 5% acidoclinata, D. smiithi v. Dilatata

MOISTURE 1- F. subsalina OCCURRING IN WATER BODIES

160

80%

20- S. OCCURRING IN WATER BODIES 45% construens, C. placentula

0- E. turgida, S. WET-MOIST PLACES, WET PLACES 10% pupula, G. affine, R abbreviata, E. adnata, S. pinnata

0- S. ND 5% acidoclinata, D. smiithi v. Dilatata

161

ZONE 2

40- F. subsalina ALKALOPHILOUS 85%

1- S. anceps CIRCUMNEUTRAL 15% pH 0- S. pupula, P. RANGING FROM ACIDOPHILOUS TO ALKALOPHILOUS 7% acrosphaeria, S. acidoclinata

40- F. subsalina BRACKISH-FRESH 85%

1- S. anceps FRESH-BRACKISH 15% SALINITY 0- S. pupula, P. FRESH, FRESH-BRACKISH 7% acrosphaeria, S. acidoclinata

40- F. subsalina TOLLERATE VERY SMALL CONCENTRATIONS OF 85% ORGANICALLY BOUND NITROGEN

1- S. anceps TOLLERATE ELEVATED CONCENTRATIONS OF 15% ORGANICALLY BOUND NITROGEN NITROGEN 0- S. pupula, P. TOLLERATE ELEVATED CONCENTRATIONS OF 7% acrosphaeria, ORGANICALLY BOUND NITROGEN S. acidoclinata

40- F. subsalina FAIRLY HIGH (ABOVE 75% SATURATION) 85% OXYGEN 1- S. anceps FAIRLY HIGH 15%

162

0- S. pupula, P. MODERATE OXYGEN 7% acrosphaeria, S. acidoclinata

40- F. subsalina α,β - MESOSAPROBOUS 85%

1- S. anceps β - MESOSAPROBOUS 15% SAPROBITY 0- S. pupula, P. OLIGOSAPROBOUS- β - MESOSAPROBOUS 7% acrosphaeria, S. acidoclinata

40- F. subsalina ND 85%

1- S. anceps MESO-EUTRAPHENTIC TROPHIC 15%

STATE 0- S. pupula, P. OLIGO-MESOTRAPHENTIC, MESO-EUTRAPHENTIC 7% acrosphaeria, S. acidoclinata

40- F. subsalina OCCURRING IN WATER BODIES 85%

1- S. anceps MAINLY OCCURRING IN WATER BODIES, SOMETIMES IN 15% WET PLACES MOISTURE 0- S.pupula, P. MAINLY OCCURRING IN WATER BODIES, ALSO RATHER 7% acrosphaeria, REGULARLY ON WET AND MOIST PLACES S. acidoclinata

163

ZONE 3 pH 20- S. pupula ACIDOBIONTIC 50%

10- C. placentula ALKALOPHILOUS 50%

0- E. adnata, G. ALKALIBIONTIC, ALKALOPHILOUS 20% affine, P. frequentisimum

0- G. truncatum, CIRCUMNEUTRAL TO ALKALOPHILOUS 10% N. viridula, N. trivialis, S. phoenicenteron, G. accuminatum

SALINITY 20- S. pupula BRACKISH-FRESH 50%

10- C. placentula FRESH-BRACKISH 50%

0- E. adnata, G. FRESH-BRACKISH 20% affine, P. frequentisimum

0- G. truncatum, FRESH-BRACKISH TO BRACKISH FRESH 10% N. viridula, N. trivialis, S. phoenicenteron, G. accuminatum

NITROGEN 20- S. pupula TOLLERATE ELEVATED CONCENTRATIONS OF 50% ORGANICALLY BOUND NITROGEN

10- C. placentula TOLLERATE ELEVATED CONCENTRATIONS OF

164

50% ORGANICALLY BOUND NITROGEN

0- E. adnata, G. TOLLERATE SMALL TO ELEVATED CONCENTRATIONS OF 20% affine, P. ORGANICALLY BOUND NITROGEN frequentisimum

0- G. truncatum, TOLLERATE SMALL TO ELEVATED CONCENTRATIONS OF 10% N. viridula, N. ORGANICALLY BOUND NITROGEN trivialis, S. phoenicenteron, G. accuminatum

OXYGEN 20- S. pupula MODERATE OXYGEN 50%

10- C. placentula MODERATE OXYGEN 50%

0- E. adnata, G. MODERATE TO CONTINUOUSLY HIGH OXYGEN 20% affine, P. frequentisimum

0- G. truncatum, MODERATE TO FAIRLY HIGH OXYGEN 10% N. viridula, N. trivialis, S. phoenicenteron, G. accuminatum

SAPROBITY 20- S. pupula β - MESOSAPROBOUS 50%

10- C. placentula β - MESOSAPROBOUS 50%

0- E. adnata, G. α – MESO/POLYSAPROBOUS- β - MESOSAPROBOUS 20% affine, P. frequentisimum

0- G. truncatum, α,β - MESOSAPROBOUS

165

10% N. viridula, N. trivialis, S. phoenicenteron, G. accuminatum

TROPHIC 20- S. pupula MESO-EUTRAPHENTIC 50% STATE 10- C. placentula EUTRAPHENTIC 50%

0- E. adnata, G. MESOTRAPHENTIC, MESO-EUTRAPHENTIC, OLIGO- 20% affine, P. EUTRAPHENTIC frequentisimum

0- G. truncatum, MESO-EUTRAPHENTIC TO EUTRAPHENTIC 10% N. viridula, N. trivialis, S. phoenicenteron, G. accuminatum

MOISTURE 20- S. pupula MAINLY OCCURRING IN WATER BODIES, ALSO RATHER 50% REGULARLY ON WET AND MOIST PLACES

10- C. placentula MAINLY OCCURRING IN WATER BODIES, SOMETIMES 50% ON WET PLACES.

0- E. adnata, G. MAINLY OCCURRING IN WATER BODIES, SOMETIMES 20% affine, P. ON WET PLACES. frequentisimum

0- G. truncatum, MAINLY OCCURRING IN WATER BODIES, SOMETIMES 10% N. viridula, N. ON WET PLACES. trivialis, S. phoenicenteron, G. accuminatum

166