POPULATION DYNAMICS OF THE AMAZONIAN PALM MAURITIA FLEXUOSA: MODEL

97 Pages·2007·1.48 MB·English

POPULATION DYNAMICS OF THE AMAZONIAN PALM Mauritia flexuosa: MODEL
DEVELOPMENT AND SIMULATION ANALYSIS

By

JENNIFER A. HOLM

A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2007
1
© 2007 Jennifer A. Holm

2
To my family who encouraged me at a young age, to keep striving for academic knowledge, and
to my friends

3
ACKNOWLEDGMENTS
I gratefully thank my supervisory committee, Dr. Kainer and Dr. Bruna, and most
importantly my committee chair, Dr. Wendell P. Cropper Jr. for their time and effort. I
acknowledge the School of Natural Resources and Conservation, the School of Forest Resources
and Conservation, and the Tropical Conservation and Development Program, the United States
Forest Service, and the Fulbright Scholar Program for funding and guidance. Data collection in
Ecuador was conducted with the help from Dr. Christopher Miller, Drs. Eduardo Asanza and
Ana Cristina Sosa, Joaquin Salazar, and all the Siona people of Cuyabeno Faunal Reserve. Data
collected in Peru was conducted with the help from Weninger Pinedo Flores, Exiles Guerra,
Gerardo Bértiz, Dr. Jim Penn, and with the generosity of Paul and Dolly Beaver of the Tahuayo
Lodge. Lastly, I would like to thank my parents for their support through my education
experience, Heather, Chris, friends, and fellow graduate students.

4
TABLE OF CONTENTS

page
ACKNOWLEDGMENTS...............................................................................................................4
LIST OF TABLES...........................................................................................................................7
LIST OF FIGURES.........................................................................................................................8
ABSTRACT...................................................................................................................................10
CHAPTER
1 INTRODUCTION..................................................................................................................12
2 POPULATION DYNAMICS OF THE DIOECIOUS AMAZONIAN PALM Mauritia
flexuosa: SIMULATION ANALYSIS OF SUSTAINABLE HARVESTING......................14
2.1 Introduction...................................................................................................................14
2.2 Methods.........................................................................................................................17
2.2.1 Study Site..........................................................................................................17
2.2.2 Study Species....................................................................................................17
2.2.3 Data Collection.................................................................................................18
2.2.4 Matrix Model Development and Parameter Estimation....................................19
2.3 Results...........................................................................................................................22
2.3.1 Density Dependence..........................................................................................22
2.3.2 Sustainable Harvest Scenarios..........................................................................23
2.4 Discussion.....................................................................................................................25
2.4.1 Sustainable Harvest Scenarios..........................................................................25
2.4.2 Implications for Management...........................................................................26
3 GENETIC ALGORIHTM OPTIMIZATION FOR DEMOGRAPHIC PARAMETER
CALIBRATION AND POPULATION TRAITS OF A HARVESTED TROPICAL
PALM.....................................................................................................................................41
3.1 Introduction...................................................................................................................41
3.1.1 South American Palms and Consequences of Wild Harvesting.......................41
3.1.2 Matrix Modeling and Population Dynamics.....................................................42
3.1.3 Parameter Calibration........................................................................................43
3.1.4 Introduction to Genetic Algorithms..................................................................44
3.1.5 Objectives..........................................................................................................45
3.2 Methods.........................................................................................................................45
3.2.1 Study Site: Ecuador...........................................................................................45
3.2.2 Study Site: Peru.................................................................................................46
3.2.3 Study Species Role in Peru...............................................................................47
3.2.4 Palm Distribution and Matrix Model Development:........................................48
3.2.5 GA Method Description....................................................................................48
5
3.3 Results...........................................................................................................................51
3.3.1 Genetic Algorithms...........................................................................................51
3.3.2 Peru Size Class Distribution and Demographic Characteristics.......................54
3.3.3 Genetic Algorithm: Harvest History for Peru M. flexuosa Palm Population....55
3.4 Discussion.....................................................................................................................57
3.4.1 Genetic Algorithms...........................................................................................57
3.4.2 Peru Size Class Distribution and Demographic Characteristics.......................59
3.4.3 Genetic Algorithm: Peru M. flexuosa Palm Population....................................60
3.5 Conclusions...................................................................................................................60
4 SUMMARY............................................................................................................................84
4.1 Applicability..................................................................................................................84
4.2 Future for M. flexuosa...................................................................................................84
4.3 Future Research.............................................................................................................85
APPENDIX: PERU GARDEN DATA FOR Mauritia flexuosa...................................................86
LIST OF REFERENCES...............................................................................................................89
BIOGRAPHICAL SKETCH.........................................................................................................96

6
LIST OF TABLES
Table page

2-1 Observed size class distribution (based on height) of M. flexuosa....................................29
2-2 M. flexuosa (Ecuador) transition matrix............................................................................30
2-3 Size class distribution for density independent (DI) and density dependent (DD)
models after 100 yr............................................................................................................31
2-4 Adult (stage 5 & 6) M. flexuosa transition probabilities....................................................32
3-1 M. flexuosa population size class distribution in Ecuador and Peru..................................62
3-2 Optimal Ecuador seedling parameters and carrying capacity using observed
demographic parameters....................................................................................................63
3-3 Observed range of the 13 non-seedling demographic transition probabilities..................64
3-4 Observed and optimal Ecuador transition matrices...........................................................65
3-5 Optimal Ecuador seedling parameters and carrying capacity using optimal
demographic parameters....................................................................................................66
3-6 Demographic traits for the Peru population in size classes 3-6 (palms that have
developed trunks)...............................................................................................................67
3-7 GA estimates of harvest regimes consistent with the observed size class distribution.....68
3-8 Peru’s harvesting history found using separate GAs which uses the Ecuador optimal
demographic parameters to reach Peru’s observed population distribution......................69

7
LIST OF FIGURES
Figure page

2-1 Elasticity for the M. flexuosa matrix population model.....................................................33
2-2 Density dependent model for M. flexuosa simulated over 500 yr.....................................34
2-3 Four simulated harvesting scenarios with various harvest frequencies and intensities
for M. flexuosa...................................................................................................................35
2-4 The average number of female palms the year before harvesting over a 100 yr
harvest regime....................................................................................................................38
2-5 Harvesting 22.45 percent at a frequency of every 20 yr....................................................39
2-6 Two harvest scenarios (both 75 percent at a frequency of every 10 yr) with density
independence (DI) and density dependence (DD).............................................................40
3-1 Map of study site in Ecuador.............................................................................................71
3-2 Map of study site in Peru...................................................................................................72
3-3 Ecuador Genetic Algorithm (GA) size class distribution and observed size class
distribution after running a GA to find optimal seedling parameters (stasis and
growth) and carrying capacity. Fitness score for this GA is 46.67....................................73
3-4 Ecuador GA size class distribution and observed size class distribution after running
a GA using the observed transition parameters and evaluating how well it matches
the observed population distribution. Fitness score for this GA is 67.03..........................74
3-5 Genetic algorithm simulated and observed size class distribution of the Ecuador palm
population for unconstrained, all values (UCAV) GA optimization. Fitness score for
this GA is 14.14.................................................................................................................75
3-6 Genetic algorithm simulated and observed size class distribution of the Ecuador palm
population for unconstrained, no two-size class transitions (UCNT) GA optimization.
Fitness score for this GA is 5.07........................................................................................76
3-7 Stasis and growth demographic points generated from a GA optimization for (A) an
UCAV (unconstrianed, all value), and (B) an UCNT (unconstrained, no two size-
class transitions).................................................................................................................77
3-8 Stasis and growth demographic points generated from CAV (constrained, all values)
GA optimization.................................................................................................................78
3-9 Stasis and growth demographic points generated from a CNT (constrained, no two-
size class transitions) GA optimization..............................................................................79
8
3-10 Genetic algorithm simulated and observed size class distribution of the Ecuador palm
population for CAV (constrained, all values) GA optimization. Fitness score for this
GA is 6.29..........................................................................................................................80
3-11 Genetic algorithm simulated and observed size class distribution of the Ecuador palm
population for CNT (constrained, no two-size class transitions) GA optimization.
Fitness score for this GA is 7.69........................................................................................81
3-12 M. flexuosa population distribution for palm populations in a 1ha area in Peru and
Ecuador..............................................................................................................................82
3-13 Distribution of male vs. female palms and estimated, averaged fecundity values from
Peru and Ecuador...............................................................................................................83
A-1 Average height(m) for M. flexuosa palms sampled in Peru (wild and gardens).
A-2 Comparison of average palm height and average number of petioles, for juvenile M.
flexuosa located in gardens and wild populations (in Peru).
A-3 M. flexuosa palms in Peru homegardens. (A) Picture of juvenile (pre-reproductive)
M. flexuosa palms in a homegardens. (B) Picture of dwarf, reproductive female palm
in Peru.

9
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
POPULATION DYNAMICS OF THE AMAZIONIAN PALM Mauritia flexuosa: MODEL
DEVELOPMENT AND SIMULATION ANALYSIS
By
Jennifer A. Holm
December 2007
Chair: Wendell P. Cropper Jr.
Major: Interdisciplinary Ecology

The tropical palm Mauritia flexuosa has high ecological and economic value, but some
wild populations are harvested excessively by cutting the stem to retrieve the fruit. It is likely
that M. flexuosa harvesting in the Amazon will continue to increase over time. I investigated the
population dynamics of this important palm, the effects of harvesting, and suggested sustainable
harvesting regimes. Data were collected from populations in the Ecuadorian Amazon that were
assumed to be stable. I used a matrix population model to calculate the density independent
asymptotic population growth rate (λ = 1.046) and to evaluate harvesting scenarios. Elasticity
analysis showed that survival (particularly in the second and fifth size class) contributes more to
the population growth rate than does growth and fecundity. In order to simulate a stable
population at carrying capacity, density dependence was incorporated and applied to the seedling
survival and growth parameters in the transition matrix. Harvesting scenarios were simulated
with the density dependent population models to predict sustainable harvesting regimes for the
dioecious palm. I simulated the removal of only female palms and showed how both sexes are
affected with harvest intensities between 15 and 75% and harvest intervals of 1 to 15 years. By
assuming a minimum female threshold, I demonstrated a continuum of sustainable harvesting
schedules for various intensities and frequencies for 100 years of harvest. Furthermore, by
10