| Abstract | Jedan od najvažnijih izvora zagaenja morskih ekosustava je onaj proizveden od strane
ljudskih aktivnosti u pripadajucim slivovima. Razumijevanje povezanosti izmeu kakvoce
morskih ekosustava i okolnih slivova važno je kako bi se bolje procijenili procesi u morskim
ekosustavima te evaluirale razlicite mogucnosti upravljanja s ciljem poboljšanja stanja morskog
ekosustava. Cilj ovog doktorskog rada je pridonijeti postizanju operativnog, na ekosustavu
baziranog upravljanja morskim ekosustavima, koji je dan u okviru Marine Strategy Framework
Directive (2008/56/EC), podizanju znanstvene spoznaje o funkcioniranju morskih ekosustava,
osobito pod djelovanjem razlicitih pritisaka iz okolnih slivova, analiza pojedinacno ili zajedno,
kako antropogenih tako i prirodnih pritisaka, kao što je obogacenje nutrijentima. Pozornost se
posvecuje procjeni odgovarajuceg stupnja procišcavanja otpadnih voda u skladu sa stanjem
morskog ekosustava.
Za podrucje istraživanja izabran je sjeverni Jadran. Sjeverni Jadran najplice je, a takoer i
jedno od najproduktivnijih podrucja u Jadranu, kao i na cijelom Mediteranu. Okolno slivno
podrucje sjevernog Jadrana rasprostire se preko cetiriju susjednih zemalja, Italije, Slovenije,
Hrvatske i Švicarske. Ovo slivno podrucje odlikuje se razlicitim pritiscima iz okoline, kako
prirodnim tako i antropogenim kao što su poljoprivreda i urbanizacija.
Da bi se postigao cilj ovog doktorskog rada korišteni se principi bazirani na DPSIR
modelu (Driving forces-Pressures-States-Impacts-Responses, odnosno Pokretne sile-Pritisci-
Stanja-Utjecaji-Odgovori), posebno definiranje Pritisaka, kako se ti Pritisci velikih morskih
ekosustava prenose na promjene Stanja, kako promjene Stanja djeluju na Utjecaje, te definiranje
odgovarajucih Odgovora.
Prvo su korištenjem Arc View Generalized Watershed Loading Function (AVGWLF)
izracunati nutrijenti (dušik-N i fosfor-P), tj. Pritisci na cijelom slivnom podrucju sjevernog
VIII
Jadrana. Drugo, upotrebom alata strojnog ucenja (Weka i Cubist) definirano je Stanje morskog
ekosustava izradom modela dinamike koncentracije fitoplanktona i modela cvjetanja mora
(TIN/PO4 model) danih u obliku regresijskih stabala. U trecem koraku povezna su Stanja morskog
ekosustava sjevernog Jadrana sa Pritiscima iz slivnog podrucja korištenjem alata strojnog ucenja.
Veza izmeu Pritisaka i Stanja temeljni je preduvjet za postizanje objektivnog, na ekosustavu
baziranog principa upravljanja morskim ekosustavima. Za ovaj zadatak korišten je Multi Target
Stepwise Model Tree Induction algoritam (MTSMOTI). MTSMOTI koristeci vrijednosti o
kolicinama ukupnog fosfora i ukupnog dušika u slivu simulira, tj. racuna klorofil, otopljeni
anorganski dušik, ukupni fosfor i zasicenje kisikom u sjevernom Jadranu preko kojih se zatim
ocjenjuju troficna stanja ekosustava korištenjem TRIX-a. Provedbom evaluacije razlicitih
scenarija koji predstavljaju razlicite opcije upravljanja tj. Odgovora u slivnom podrucju (npr.
razliciti stupnjevi procišcavanja otpadnih voda, gnojivo s manjom koncentracijom nutrijenta
poljoprivredi i sl.) te njihovim utjecajem na morski ekosustav (TRIX) definirani su odgovarajuci
stupnjevi procišcavanja otpadnih voda, što predstavlja znacajan Odgovor kod smanjenja kolicine
nutrijenata. |
| Abstract (english) | One of the most important sources of pollution in marine ecosystems are those produced
by human activities in the associated watersheds. Understanding the linkage between water
quality of marine ecosystems and surrounding watersheds is important in order to better assess
processes in marine ecosystems and to evaluate different management options aimed at improving
the marine ecosystem state. The goal of this Ph.D Thesis is to contribute to achieve an operational
ecosystem-based management of marine ecosystems, which has been called for in the Marine
Strategy Framework Directive (2008/56/EC), by improving the scientific knowledge of the
functioning of marine ecosystems, especially under the actions of different pressures (such as
nutrient enrichment) from surrounding watersheds, analyzed alone or together, both
anthropogenic and natural. Attention is paid to the assessment of proper wastewater treatment
level according to the desired marine ecosystem state.
Northern Adriatic (NA) is chosen for the case study area. NA is the shallowest and
subsequently one of the most productive parts in the Adriatic, as well as in the whole
Mediterranean Sea. Surrounding watershed area of NA is spread over four neighboring countries,
e.g. Italy, Slovenia, Croatia and Switzerland. This watershed is characterized by different natural
as well as anthropogenic pressures, e.g. agriculture and urbanization.
To achieve the goal of this PhD Thesis the principles of Driving forces-Pressures-States-
Impacts-Responses (DPSIR) framework were used, in particular on defining Pressures, how
Pressures exerted on large-scale marine ecosystems translate into State changes, how State
changes act to Impacts and to define corresponding Responses.
First, ArcView Generalized Watershed Loading Function (AVGWLF) was used to
calculate nutrient loadings e.g. Pressures (nitrogen and phosphorus) from surrounding NA
watershed. Second, using machine learning (ML) tools (Weka and Cubist) State of marine ecosystem was defined trough development of phytoplankton and mucilage descriptive models in
a form of regression trees. In the third step the State of NA marine ecosystem was linked to
Pressures from surrounding watershed also with the use of ML tools. The link between Pressures
and States is a fundamental prerequisite to achieve a real ecosystem-based approach of marine
ecosystems. For this task was used Multi Target Stepwise Model Tree Induction (MTSMOTI).
MTSMOTI simulates chlorophyll-a (Chl-a), dissolved inorganic nitrogen (DIN), total phosphorus
(Ptot) and oxygen saturation (Osat) in NA from which trophic conditions of the ecosystem are
being evaluated using the TRIX. Through evaluation of various scenarios which present different
management options e.g. responses in the watershed (e.g. different wastewater treatment level,
fertilizers with less nutrients in agriculture etc.) and through their influence on marine ecosystem
(TRIX ) the proper level of wastewater treatment was determined which is significant response in
reduction of nutrients. |
