ECOSYSTEM SERVICES IN HUNGARIAN KARST AREAS, WITH A BAYESIAN NETWORK MODEL OF EUTROPHICATION

kettlecatelbowcornerAI and Robotics

Nov 7, 2013 (4 years ago)

85 views

Salzau 2010



Abstracts

February
2010








Department of Cllimatology and Landscape Ecology

University of Szeged, Hungary



Karst areas are among the most valuable landscapes of Hungary, and they are also
very vulnerable. Karstic rocks (limestone, dolomite) have good transmissibility and
water
holding capacity, therefore specific surface and sub
-
surface geomorphologic
forms can be found on them; the higher geodiversity results in the diversity of living
organisms. But the rocks’ good solubility, the three
-
dimensional processes are a
source of ha
zards as well, because harmful effects quickly induce significant
changes in the whole system. Besides water (with transmitter function), every
landscape ecological factor has an important role in the system, and they affect the
special morphology and proc
esses of the karsts, thus we proposed the concept of
karstecosystem
. We would like to develop this concept by evaluating ecosystem
services.


Our case study deals with the problem of eutrophication of karstic lakes. The nutrient
input originating from hum
an activity has led to a decline in water quality, which was
indicated firstly by losing some functions of these water bodies, and ultimately led to
their role in the ecosystem completely disappearing. Some of these lakes have lost
almost their entire unco
vered water surface within decades. The complex causal
connections of the process of eutrophication can be modelled efficiently with
Bayesian Networks, which we used based on cases with measured data. The first
part of our model demonstrates one possible c
onsequence (proportion of poisonous
ammonia, which is poisonous to fish populations) of the eutrophic state caused by
nutrient overload, and the factors that affect it (water temperature, pH, nutrient input,
chlorophyll
-
a, alkalinity, SPI drought index). T
he second part estimates economic
value of decreasing fish populations with direct use valuation methods. Based on
these initial results, we regard Bayesian Networks as usable method for evaluating
wetland ecosystems.





ECOSYSTEM SERVICES IN HUNGARIAN KARST AREAS, WITH A
BAYESIAN NETWORK MODEL OF EUTROPHICA
TION


Márton Kiss, Andrea Samu, Eszter Tanács, Ilona Bárány
-
Kevei