Spatiotemporal analysis and modeling of ecological processes at ecosystem, landscape and bioregion scale
Author(s)
Ciolfi, Marco
Date Issued
June 27, 2016
Type
Doctoral Thesis
Abstract
The ecological systems are complex and multifold: many variables cooperate in the definition of
the measured values and countless disturbances affect the sampling. Furthermore, often ecolog-
ical systems show patterns of change both in time and space. Complexity calls for compromise:
choosing a single driver of change, be it space or time, is always somewhat arbitrary owing
to pre-packaged statistical analysis tools. Geographical Information Systems (GIS) are now
widespread and offer many spatial modeling tools, but they lack the ability to capture the time
variability. The statistical analysis of time series, on the other hand, is blind to spatial patterns.
This work proposes an interpolation tool, which is able to include spatial as well as tem-
poral variability: the Timescape Algorithm. This algorithm derives form the fusion of the
consolidated spacetime interpolation techniques of Statistical Physics with the typical needs of
ecological systems modeling. Two software versions have been derived from the general algo-
rithm: a local (projected coordinates) and a global (longitude/latitude) version. The software
is published according to an open license (GNU-GPL v3.0) and is distributed with a detailed
manual, the source code and a sample dataset of isotopic abundances in precipitations. The
software has been designed to be the slightest possible detour from a consolidated GIS work
ow,
since more and more researchers use a GIS environment for both data storage and modeling.
The Timescape idea is the mathematical translation of a few quite general statements:
-observations at different places and times can be mixed freely
-the area of possible in
uence grows with time
-closer sites can in
uence each other more than far ones
-the value at a given time can in
uence the future values
The Timescape Algorithm has been presented at the 2016 European Geoscience Union Conference
[Geophysical Research Abstracts Vol. 18, EGU2016-15864, 2016].
Spatial and spatiotemporal modeling is illustrated through three case studies, re
ecting the
author's activities within his grant at the Institute of Agro-Environmental and Forest Biology
of the National Research Council of Italy.
The first case study, Mycorrhiza Survival Strategy, is centered on the use of carbon and
nitrogen stable isotopes relative abundances to investigate the symbiotic relationships among
1
fungi (Tuber aestivum) and host trees. According to the current literature, the interaction has
been modeled according to a simple isotope fractionation scheme: carbon
owing from host to
fungus shows negligible fractionation whilst nitrogen
owing from mycorrhiza to trees displays
fractionation. This research was the main driver for the development of the Timescape idea,
since the temporal variability could not be neglected for the truffles, which can continuously
change the isotopic compositions of the fruit bodies during the collection season. The
experiment was conducted on a protected area in Umbria region, within the STINA interregional
park. Here, old Pinus spp. stands were partially replaced with broadleaved species. Several
detailed Isoscapes (thematic maps that show the spatial distribution of stable isotopes, thus
tracing an isotopic landscape) were produced in order to map the spatial patterns of soil 15N
and leaves 15N and 13C. Soil and leaves did not need any special statistical treatment because
their isotopic content is accumulated over time (one season for the leaves and longer times for
the soil). Stumps were also sampled.
The modeling shows a remarkable probability of symbiosis among tru es and pines. No
signi cant statistical matching has been found with other species in the pine-cleared areas,
suggesting a saprophytic behaviuor of the mycorrhiza, feeding on the remnant stumps. This
relation seems to be con rmed by the isotopic signatures of the examined stumps.
The results of this study have been presented at the First Isotope Ratio Mass Spectrometry Day -
Constitution of the Italian IRMS Group [F. Camin (ed) - Proc 1st IRMS Day, ISBN-978-88-7843-046-4].
The second case is an Extra Virgin Olive Oil (EVOO) provenance assessment study
for the geographic origin through carbon (13C) and oxygen (18O) stable isotopes. The aim is
the protection of consumers from geofrauds i.e. the false declaration of origin which raise the
product price pretending to use olives only from highly-reputed geographic areas. This kind
of fraud is not related to the adulteration of EVOOs, for which consolidated chemical analysis
methods exist, but to the subtler eld of falsely claimed geographical provenance.
This study takes into account the isotopic compositions of 387 EVOO samples originated
from di erent regions in the years of the 2009, 2010 and 2011 seasons. The autenticity of
each harvesting was certi ed by UNAPROL, the consortium of EVOO producers. Ancillary
meteorological and geomorphological data have been imported in a GIS framework for spatial
isotopic modeling.
In this case, contrary to the previous case study, there was no need to analyse time and space
variability in a single model, since every year was treated independently from the others. This
2
is sensible since, statistically, both 13C and 18O fatty acids content integrates the carbon and
water uptake from the atmosphere and available water, through the photosynthetic assimilation
and the fractionation process occurring during the transpiration. It was assumed that the
available water essentially re
ects the integrated isotopic composition precipitations.
The 13C and 18O values were compared to some geographical and meteorological param-
eters. The best explanatory variables for isotope composition variations have been found to
be the precipitation 18O and the xerothermic index, with various R2 gures according to the
year; con dence intervals were calculated too. The result is a set of predicted isotopic maps
(Isoscapes): predicted 13C for 2010 and 2011, and predicted 18O for 2009, 2010 and 2011.
The study of the predicted Isoscapes highlighted four zones: North, Central Thyrrenean,
Southern Adriatic and Islands, which of course include more than one actual DOP and IGP
production area. Isoscape analysis is a promising technique for contrasting geofrauds, especially
if supported by other chemical analyses (e.g., heavy elements content).
This study is published on Food Chemistry [Chiocchini et al. - Food Chemistry 202 (2016) 291-
301. doi:10.1016/j.foodchem.2016.01.146].
The last case study concerns an ongoing project: Ecua
ux. It is an international project
about carbon
ows and photosynthesis responses of Polylepis reticulata (locally called \paper
tree" for the nely layered structure of the bark), an endemic high altitude tree of the Southern
Ecuadorian Andes. The collection of samples, including soil, tree cores, leaves and other tissues,
has started in January 2016 and is planned to cover three or four years.
The collection is being conducted on six separate forest plots in three di erent catchments
at about 4000m altitude, at the utmost altitudinal limit of the species. The sampling has been
carefully designed for space and time modeling and will provide the ideal playground for testing
the Timescape algorithm.
The isotopic measurements (estimated number of about 2000 samples) will be centred on
the reconstruction of the changes of the photosynthesis response in the area over the last years.
Indeed, the study's site, the Cajas National Park in Ecuador, is an hot spot of climate change,
being located at such high altitude on the equator and exposed to the climatic in
uence of the
Amazon Basin and the Paci c ocean as well.
A few preliminary results can be seen on http://www.ub.edu/ecologia/ecua
ux.
Additional information
Dottorato di ricerca in Scienze e tecnologie per la gestione forestale e ambientale
File(s)![Thumbnail Image]()
Name
mciolfi_tesid.pdf
Size
52.08 MB
Format
Adobe PDF
Checksum (MD5)
04bec6cfdfb9cdb0b17ec2c798b9458b