Master’s thesis
Physical Geography and Quaternary Geology, 45 Credits
and Quaternary Geology
Local water resource
assessment in Messinia, Greece
Karin Ekstedt
NKA 81 2013
Preface
This Master’s thesis is Karin Ekstedt’s degree project in Physical Geography and Quaternary Geology at the Department of Physical Geography and Quaternary Geology, Stockholm University. The Master’s thesis comprises 45 credits (one and a half term of full-time studies).
Supervisors have been Jerker Jarsjö and Steve Lyon at the Department of Physical Geography and Quaternary Geology, Stockholm University. Examiner has been Karin Holmgren at the Department of Physical Geography and Quaternary Geology, Stockholm University.
The author is responsible for the contents of this thesis.
Stockholm, 13 June 2013
Lars-Ove Westerberg Director of studies
1
ABSTRACT
Messinia is a region in Greece renowned for its rich nature, olive agriculture and water availability. In the light of increasing anthropogenic and climatic pressure, this study assessed local water resources in catchments in south western parts of the region. The main objectives were to evaluate the balance between supply and demand, the sustainability of current water consumption, capacity of further land use intensification and to review local water
management. The method was dual with both quantitative (water balance calculations and linear modeling) and qualitative (interviews and a questionnaire survey) approaches.
It was confirmed that, on an annual basis, rainfall is comparatively high, there is a surplus of water leaving the catchments and aquifers are “superfluous”. The climate however, brings seasonal imbalance and notable shortages during summer that affect operation of local actors, especially with agriculture and tourism being the principal water users. Unofficial sources indicated that current consumption may not be sustainable, either because of over-exploitation or climatic changes, but further studies are required to draw reliable conclusions. Modeling showed the importance of land management, that unconsidered water consumption may impact the water balance substantially but also that, while minimizing evapotranspiration, there is capacity of intensification if water withdrawals are increased. Considering
accessibility, competitive interests and sustainability however, such development is not necessarily feasible.
The municipal water management appeared to be well established and, given that measures are taken concerning for example stakeholder integration and regulation of private and agricultural consumption, there is capacity of handling increasing water stress. Finally, stressing the crucial role of freshwater availability, the study highlighted the importance of further hydrological research and thus the need for improved data quality, particularly regarding river discharge.
ACKNOWLEDGEMENTS
First of all, I would like to give big thanks to Jerker Jarsjö and Steve Lyon for more than great support and supervision throughout the project. Big thanks go also to the NEO (see below) station managers Nikos Kalivitis (up to July 2012) and Giorgos Maneas who have been really helpful with a variety of issues. Konstantine Boulolis, agronomist in Gargaliani, also deserves special thanks for all the valuable information and material provided; during the interview and in later e-mail conversations. The support from Efstathia Zontanou at Nileas, with the
interview and with collecting questionnaires is likewise highly appreciated. I am grateful also to Victoras Plevrakis for consultation and pictures and, together with Katerina Mazi and Magdalini Zampouni, for translations to and from Greek. I furthermore want to mention Göran Alm and Ingmar Borgström at the department who helped with lending a GPS and providing data and material respectively. Thanks go also to Iris Claesson, at the Department of Human Geography, for lending the Dictaphone. Finally, I want to thank TEMES for providing the data, all that I interviewed and those who filled in the questionnaires and, of course, family and friends that supported and encouraged me along the way. No report without all of you.
2
3
TABLE OF CONTENTS
1. INTRODUCTION ... 5
1.1 Study objectives ... 5
1.2 Regional context ... 5
2. DATA AND METHODS ... 10
2.1 Site description ... 10
2.2 Data ... 12
2.3 Methods ... 16
3. RESULTS ... 22
3.1 The water balances ... 22
3.2 ET of local vegetation and olive irrigation ... 23
3.3 Hypothetical land use intensification ... 23
3.4 The interviews and questionnaires ... 25
4. DISCUSSION ... 31
4.1 Performance and uncertainty ... 32
4.2 The current state of local water resources ... 33
4.3 Local water management ... 37
4.4 Further studies ... 39
5. SUMMARY AND CONCLUSIONS ... 40
REFERENCES ... 41
APPENDIX A: STUDIES REVIEWED FOR ESTIMATING ET IN OLV AND NV ... 46
APPENDIX B: THE QUESTIONNAIRE ... 48
APPENDIX C: RESULTS OF MODEL 1 IN GIANOUZAGAS ... 50
APPENDIX D: COMMENTS TO THE QUESTIONNAIRS ... 51
4
5
1. INTRODUCTION
Freshwater availability is integral to all ecosystems and to all aspects in human societies, economic as well as social. Therefore shortage can be a major constraint to development. In the semi-arid Mediterranean region water resources are indeed limited, especially because of the characteristic of seasonal imbalance in supply and demand that induces scarcity in the summer months. This is also an area where climate change is projected to decrease water availability further and where anthropogenic pressures are increasing through growing population, tourism and irrigation demands (e.g. Tsagarakis, et al., 2003; IPCC, 2007a; b;
Trondalen, 2009).
Thus, understanding and monitoring the hydrological system and the anthropogenic influence on it, is of high importance in both land and water management in this region (Sánchez- Canales, et al., 2012). Research is fundamental on all scales and there is a need for
coordination and cooperation, between scientists and with stakeholders and policy makers.
(Cudennec, et al., 2007). From such motivation, this thesis looks at local water resources and management in the area round Navarino Environmental Observatory (NEO), Greece, that is a recently started cooperation between Stockholm University, the Academy of Athens and TEMES S.A. The NEO is located at Costa Navarino, a luxury mixed-use resort being developed by TEMES, and is meant to gather researchers from all over the world to offer a platform for studies and knowledge exchange on Mediterranean environments.
1.1 Study objectives
This study’s main objective is to evaluate the current state of water resources in three local catchments in the area of NEO: the Sellas, Gianouzagas and Xerias rivers (see site
description). The study addresses both water availability and demands and an attempt is made at both reviewing local water resources management and at assessing the situation and
opinion of local actors who deal with, and are depending on, the local water resources. More specifically, it is investigated i) if there is shortage or surplus of water resources, ii) if current levels of consumption are sustainable, iii) how local water management is set up and iv) if management is sufficient to address increasing anthropogenic and climatic pressure in the area. Also, to address current and future development, the capacity for hypothetical land use intensification is simulated through simple linear modeling with emphasis on how such intensification would affect local hydrological systems.
The overall method includes setting up the catchment water balances and modeling, but also the gathering of information from local stakeholders through interviews and questionnaires during a long-term stay on location. This dual approach offers a great opportunity of
characterizing local conditions and of identifying up-to-date key points. The hope is that the thesis will provide an inclusive overview of the local hydrological system and make a starting point for further water or other environmentally related studies in the area.
1.2 Regional context
Costa Navarino is located in the western parts of Messinia in southwestern Peloponnese, Greece (see site description), an area renowned for its rich nature and prosperous olive agriculture in which human settlement reaches far back in time. The geography of the region
6
was described in the 19th century already and a number of physical studies have followed, often in connection to archaeological investigations such as the “Minnesota Messina Expedition” (McDonald and Rapp, 1972) and the “Pylos Regional Archaeological Project”
1991-1996 (Zangger, et al., 1997). Investigating local hydrology requires an understanding of the regional context. Therefore the following sections offer a more detailed presentation of principal regional parameters: climate, climate change, land use and water management.
1.2.1 Regional climate and water resources
For being under Mediterranean climate, weather in Messinia is relatively humid. Most precipitation (P) in Greece is brought in by westerlies during winter and Messinia, being situated on the west coast and on the windward side of the inland mountains, receives relatively large amounts of rainfall. Annual P averages 1100 mm in this part of Greece but varies over an elevation gradient starting at circa 800 mm at the coast, reaching 1000 mm further uphill and around 1600 mm in the higher mountains (Loy and Wright, 1972; Ministry of Development, 1997). This can be compared to an average of 652 mm over the entirety of Greece (AQUASTAT, 2013) and to 350 mm round Athens and on the east coast - where the land is shadowed by the Pindos mountain range extending north-south through the country (Loy and Wright, 1972; Ministry of Development, 1997; Baltas, 2008).
Despite relatively large amounts, rainfall is still characterized by the Mediterranean
seasonality which is controlled by the north- and southward shift of the large-scale general circulation systems. In winter, moist air and cyclonic depressions are brought in from the Atlantic by the Subpolar jet stream and the westerly surface winds, while in summer, a northward shift places the Bermuda-Azores subtropical high pressure system over the region.
Thus in summer, i.e. June-August, descending and warming air dominates, skies are clear and P amounts are close to zero. About half the annual P falls in winter December-February, mostly as rain, and the other half in autumn and spring. Summer thunderstorms are rare in the region but do occasionally break the summer drought (Loy and Wright, 1972; Giorgi and Lionello, 2008; Finne, et al., 2011).
Temperature (T) is also bound by the seasonality of the region. Mean T in Messinia is around 11°C in winter and 27°C in summer (Loy and Wright, 1972). Generally along the coasts in Greece, mean minimum T in January-February is 5-10°C and mean maximum T in July- August is 29-35°C. Sea breezes cool temperatures along the coasts during summer (HNMS, 2013) yet the intensive incoming solar radiation still causes high potential evapotranspiration (PET) and atmospheric water demands that clearly exceed P (Fig. 1, after Loy and Wright, 1972). Hence, through the 5-6 months that make up the dry season, roughly April/May- September/October, evapotranspiration (ET) is controlled by P in combination with soil moisture and the system is clearly water-limited. Winters offer the opposite, P is rich and low temperature gives low PET (e.g. Loy and Wright, 1972; Baltas, 2008; Ryu, et al., 2008;
HNMS, 2013).
7
P PET ET
1
2 3
200
150
100
50
Figure 1: Constructed graph (fictive numbers) generalizing the typical annual interplay between P, PET and ET under Mediterranean conditions. At point 1 water surplus is ended and PET exceeds P, ET draws on soil moisture. At point 2 soil moisture is depleted and there is a clear water deficit. At point 3 P exceeds PET again and soil moisture is recharged. After original water-balance diagrams from the local area produced by Loy and Wright (1972), p.39, Fig. 3-1.
Water availability in Greece follows the complex pattern of P and the country is split into several smaller hydrological regions, ranging from those of intense deficiencies to those of lasting surpluses (Sofios, et al., 2007; MEECC, 2010). As mentioned, P and therefore also water resources, are generally richer in the western regions (Kerkides, et al., 1996) and according to the Ministry of Development (1997), the district of Western Peloponnese has ample surface- and ground water resources. Ground water flows in particular, are important for water transport in this type of climate and landscape (Newman, et al., 1998; 2006). The geological structure of Messinia, with limestone, sandstone and conglomerates, is favorable for ground water generation and the mountain headwaters make up good catch basins (Loy and Wright, 1972; García-Ruiz, et al., 2011). As a result, there are numerous springs at lower altitudes and ground water can be easily extracted in wells. This is important for feeding the extensive irrigation systems as well as for the drinking water supply, particularly in the summer drought when most rivers run dry (Loy and Wright, 1972, MEECC, 2010).
1.2.2 Climate change and water resources
Numerous studies have reviewed the characteristics and impacts of climate change in the Mediterranean region. It is identified as one of the primary “Hot-Spots” (Giorgi, 2006) and most responsive and vulnerable regions to climate change globally (e.g. IPCC, 2007a; Giorgi and Lionello, 2008; Trondalen, 2009; Bosello, et al., 2013). Greece moreover, is among the areas identified as particularly affected in the Mediterranean region (Diffenbaugh, et al., 2007). Substantial climate changes are projected with evident consistency in most models and recent studies show that shifts are already occurring (e.g. Kostopoulou and Jones, 2005;
IPCC, 2007a; b; Mavromatis and Stathis, 2011).
In short, with global warming a “pole-ward extension” is expected of the seasonal shifts in latitudes of the global circulation system. This means the Bermuda-Azores anticyclonic cell will become more dominant and long-standing in the region. The dry summer season will thus be prolonged and intensified and winter low pressures hampered and moderated (IPCC, 2007a; Giorgi and Lionello, 2008). Accordingly, P is confidently projected to decrease, especially during summer and in terms of the number of rainy days in winter. IPCC (2007a)
8
for example, predicts a total decrease in average P of 7-27% until 2090-2099 in the
Mediterranean (assembled models, SRES scenario A1B, reference period 1980-1999). The corresponding modeling for T projects an increase of 2.2-5.1°C and particular warming in summer. Many studies also point to higher rainfall intensity, higher interannual variability in both P and T and a higher frequency and severity of extreme hot and dry weather conditions (e.g. Diffenbaugh, et al., 2007; IPCC, 2007a; Trondalen, 2009; MEECC, 2010; Sánchez- Canales, et al., 2012).
Naturally then, most models predict a decrease in runoff and ground water infiltration in the Mediterranean basin (Trondalen, 2009; MEECC, 2010). This has, again, been confirmed in several studies to occur already (e.g. García-Ruiz, et al., 2011; Mavromatis and Stathis, 2011).
IPCC (2007b) predicts a decrease in runoff of 0-23% until 2020 and of 6-36% until 2070 in southern Europe (reference period 1961-1990). Together with increased interannual
variability, and a higher frequency and severity of droughts, this will cause notable
disturbances in hydrological systems altering the regime, quantity, quality and sustainability of water resources. Environmental and social impacts extend also to for example forest fires, losses of biodiversity, desertification and also effects in general health, tourism and energy consumption (e.g. IPCC, 2007b; Trondalen, 2009; FAO, 2011; García-Ruiz, et al., 2011).
1.2.3. Land use and water resources
Both natural and anthropogenic changes in land cover and land use are also particularly intense in the Mediterranean landscape, as they have been for the past 10 000 years (García- Ruiz, et al., 2011), and many studies have shown related impact in the hydrological systems.
Particularly, vegetation has large influence on evapotranspiration and infiltration that affects the quantity, quality and regime of rivers and aquifers (e.g. Brown, et al., 2005; Bhattarai, et al., 2008; García-Ruiz, et al., 2011; Sánchez-Canales, et al., 2012). The actual influence will always depend on specific catchment characteristics, such as soil water storage capacity and climatic conditions, and new equilibriums can take many years to reach after permanent changes occur (Bosch and Hewlett, 1982; Brown, et al., 2005).
In the Mediterranean, climatically induced water scarcity is projected to cause considerable changes in land use and specifically, a decrease and degradation of arable cropland is expected (e.g. MEECC, 2010; IPCC, 2007b; FAO, 2011; García-Ruiz, et al., 2011).
Agriculture has always been the main feature in Greek economy (Loy and Wright, 1972).
Sugar beets and olives dominate production and, after Spain and Italy, Greece is the third largest producer of olives in the world (FAOSTAT, 2011). Due to the climatic conditions, irrigation in Greece already accounts for 84% of total water consumption (MEECC, 2010) and demands are increasing, partly also following growing markets for olive products and
intensification in agriculture (e.g. Metzidakis, et al., 2008; Iniesta, et al., 2009; Nainggolan, et al., 2012). In Greece alone, the proportion of irrigated fields increased 22% 1990-2006
(MEECC, 2010).
Olive trees are resistant to drought and have long been cultivated in low density orchards using rainfall only. New orchards however, are streamlined with higher density, less alternate bearing behavior and effective drip irrigation (Beede and Goldhamer, 1994). Increasing water
9
demand however, poses an issue in water management and is a limiting factor for alternative uses (Bosello, et al., 2013). Research is becoming progressively focused on estimating and modeling specific and local irrigation requirements and it is crucial for sustainability that land use and irrigation management is adapted accordingly (e.g. Orgaz, et al. 2006; Testi, et al., 2006; Metzidakis, et al., 2008).
Coastal regions in the Mediterranean are under additional pressure from for example tourism (which just like agriculture is an important feature in Mediterranean economies), denser population and industry and, in addition, future scenarios indicate further intensification.
Tourism and agriculture in particular, cause seasonal imbalance in supply and demand that above all generates large deficiencies in summer. It is not uncommon that aquifers here are intensively and unsustainably exploited, not least along Greek coastlines, and that the issue of salt water intrusion emerges (e.g. Cudennec, et al., 2007; MEECC, 2010; García-Ruiz, et al., 2011; Mazi, et al., 2013). Increasing salinity, together with intense fertilization and use of pesticides in agriculture, deteriorates water quality and affects all users in these areas (NTUA, 2007; Baltas, 2008).
1.2.4 Water management in Greece
There have been recent large-scale reorganizations in Greece restructuring a fragmented and decentralized institutional framework with the merging of municipalities into larger units as one of the main outcomes. Up until 2011 the former Ministry of Environment, Planning and Public Works was the main body for environmental and water management (Tsagarakis, et al., 2003) yet in the new system, water management is governed under the National Water Commission chaired by the Ministry of Environment, Energy and Climate Change (MEECC).
The Special Secretariat for Water is then the body responsible for planning and coordinating implementation between national and regional levels (MEECC, 2009).
There are 14 Water Regions in the country, with their own departments for water and waste water, that are responsible for regional implementation of the national strategic planning (Tsagarakis, et al., 2003; MEECC, 2009). Actual measures and management however is run at municipal level. Most commonly, in municipalities with more than 10 000 inhabitants (and in some with fewer), water is managed by the “Municipal Enterprise for Water Supply and Sewage” (DEYA). There are now more than 200 DEYA in Greece and they serve about 35- 40% of the total population. The bigger cities of Athens and Thessaloniki have similar yet somewhat different solutions and smaller municipalities cover management themselves (Tsagarakis, et al., 2003; Safarikas, et al., 2006).
The DEYA is owned by the municipality and partly financed by the state yet it is run as a private company rendering the enterprise flexible and efficient in its operation (Tsagarakis, et al., 2003; Safarikas, et al., 2006). It owns the facilities, sets the water tariffs according to operation costs and independently determines their extent of cooperation with private
companies. In effect it is responsible for constructing, maintaining and running local networks for water supply and sewage and for verifying that quality is sufficient for environmental and health requirements (Tsagarakis, et al., 2003).
10
The DEYAs in Greece typically face several challenges, for example with loans, distributing networks and sewage systems requiring expensive upgrades, inadequate monitoring,
increasing anthropogenic development and increasing quality issues in local water resources (Tsagarakis, et al., 2003). In this context there are several studies that point to the importance of strong institutional frameworks with political and economic stability, fewer water districts (preferably at watershed-level as is set in the EU Water Framework Directive (WFD)) and clear guidelines and assignment of responsibilities. Also commonly emphasized is the significance of stakeholder integration; for information-sharing, for prevention of mistrust, conflict and unawareness and for fair decision-making. This is particularly important considering external political regulation commonly interferes with economic interests (e.g.
IPCC, 2007b; Baltas, 2008; García-Ruiz, et al., 2011; Bosello, et al., 2013). As mentioned, a new legal framework is being developed in Greece in order to facilitate the enterprises and water legislation is under continuous adjustment in correlation with the EU WFD (Tsagarakis, et al., 2003) - which was the focus of the first meeting of the National Water Commission in 2010 (MEECC, 2009).
2. DATA AND METHODS 2.1 Site description
Sellas, Gianouzagas and Xerias are three main rivers draining into the sea along the coast between Costa Navarino and Pylos. Their headwaters are typically in the inland mountainous areas where elevations reach more than 1000 m (Tab. 1a) but closer to the coast topography levels out and the rivers flow through plains of fertile agriculture before reaching the sea (Fig.
2). The Sellas catchment is roughly twice the size of the Gianouzagas and Xerias catchments and it also covers higher altitudes than does the other two.
The dominating land use in this area is olive agriculture (OLV). It covers around 70% of the total catchment area in Sellas and Xerias and almost 90% in Gianouzagas (Tab. 1b). Other vegetation, summed under “native vegetation” (NV), is diverse and includes coniferous, broadleaved as well as sclerophyllous vegetation. Sparsely vegetated areas and shrubs are also common but bare land and artificial surfaces cover no more than 1.5% in either of the
catchments (Lundholm, et al., 2009). Typically, geology is mainly made up of sandstone and limestone while alluvial deposits dominate at the coastal plain between Gialova and
Romanos.
Table 1: Physical characteristics of the catchments: a. altitude statistics: average (avg) (also in Tab. 2), standard deviation (std), minimum (min) and maximum (max) and b. areas: the total catchment, olive agriculture (OLV) and “native” vegetation (NV).
a. Altitude statistics, masl b. Areas, km2 (%) Avg std min max total OLV NV Sellas 370 194 6 1041 87.7 65.2 (74) 22.6 (26) Gianouzagas 268 100 20 703 39.0 34.7 (89) 4.4 (11)
Xerias 201 96 9 823 47.7 33.6 (70) 14.1 (30)
11
Figure 2: The three catchments, the local rainfall (P) and discharge (Q) gauging stations, local
villages, the two development sites of Costa Navarino, land use (Lundholm, et al., 2009) and elevation relief (Delrue, unpubl.).
At this alluvial plain lies Gialova (Osmanaga) lagoon which is perhaps the most prominent environmental feature in the area. It is an important wetland for more than 271 species of residential and migratory birds and it is the only European residence for the African
chameleon (TEMES S.A., 2009a). Through the past decades however, it has been under much anthropogenic pressure from for example freshwater withdrawals in agriculture, exhaustive fishery, drainage and increasing infrastructure (HOS, 2013). The lagoon was formed gradually
Coordinate System: WGS 1984 Projection: Mercator Auxiliary Sphere
Land useand use Land use
12
through the successive steps of barrier formations that shaped Navarino Bay the last 9000 years. Today the floodplain is an inactive erosional environment but it was previously fed with massive amounts of sediments from the Sellas River. The river was diverted just north of Romanos during the Late Bronze Age (in the Mycenaean Era, 1600 BC - 1100 BC) in a hydraulic project that created a clean water port for the palace of Nestor - the first known artificial harbor installation of its size in Europe (Zangger, et al., 1997).
The area has been of high significance and both a cultural and demographic center through history yet at present, it is a remote and rural area and population density is low (Zangger, et al., 1997). NEO and Costa Navarino are located in the municipality of Pylos-Nestor, formed from merging 6 former units in 2011, that has a population of 21 077 (38 people per km2, total area 555 km2) and in which the main city is Pylos with 2 350 residents (ELSTAT, 2012). The municipality has an oblong shape stretching from Palea Vrisi in the north to Koroni in
southeast. All in all, the area has high environmental and recreational values that make it an attractive destination for tourism. Hitherto there has still only been little tourism yet the recent establishment of Costa Navarino could make a turning point for such development. The resort is likely to attract not only thousands of guests but also further local business and settlement that will increase anthropogenic pressure on natural resources, such as water.
2.2 Data
The water balances for the three catchments (see method section) were set up using local discharge (Q) and P data series measured and provided by TEMES through NEO. They were reported every 15 minutes and the longest records covered January 2009 to October 2012 (Tab. 2). There were six meteorological stations around the catchments (Fig. 2) that were set up in a network by TEMES in 2008 to monitor for example T, humidity and wind speed. P was gauged using weighing buckets and both missing data and outliers did occur in the series.
Q was measured in the rivers of Sellas, Gianouzagas and Xerias respectively (Fig. 2 and 3) using an optical radar technique (Fig. 4). It was reported as an “instrument to water surface distance” that was converted to a depth and rated into discharge. However, no explicit
information on the “instrument to river bed distance” was provided such that it was necessary to assume it equal the highest “instrument to water surface distance” reported.
Under an initial quality control check the daily streamflow data series from Sellas appeared to be the most reliable among the three. It had the most continuous record and a hydrograph reflecting the expected river regime (Fig. 3). Gianouzagas on the other hand had very short time coverage and its hydrograph, particularly considering i) the unexpectedly steady flows in November 2010 - February 2012 and ii) the peculiar drop in March and April 2012, indicates that the measurements were affected not only by natural factors and that there were errors inherent in the observations.
13
Table 2: P and Q gauges set up by TEMES, their station names, time coverage, mean altitude and average annual P or runoff (R, see below) respectively. Brackets in the last column give Q in m3/s.
Parameter Station Time period Altitude, m P/R (Q), mm/yr
P
Navarino Sep10 – May12 34 694
Moyzaki Jan09 – May12 436 1151
Chora Jan09 – May12 237 766
Handrinos Jan09 – Oct12 184 895
Sgrapa (Gialova) Jan09 – Oct12 14 660
Kynigos Jan09 – Oct12 311 880
Q
Sellas Jan09 – May12 370 378 (1.05)
Gianouzagas Jan09 – Apr12 268 392 (0.48)
Xerias Oct10 – May12 201 1602 (2.42)
The Xerias data series contained much missing data, several unaccountable extremes (sometimes exceeding 70 m3/s) and reported values of questionable accuracy (see July- October 2009 for example). For relevant presentation in Fig. 3, extremes have been cut and missing data replaced by monthly averages. Because of data shortage and lack of reliability in the records of Xerias, these monthly averages were estimated through correlation with Sellas:
the proportional relation between the monthly and the annual average “instrument to water surface distance” in Sellas was applied to the annual average in Xerias thus giving an
approximation of the corresponding monthly values. The correlation was justified given i) the spatial proximity, ii) the assumption of similar hydrological regimes and iii) the statistical similarity of the monthly average measurements in the two catchments despite deviation some months (the annual average/standard deviation of the “instrument to water surface distance” is 302/11 and 366/16 mm in Sellas and Xerias respectively).
It should be mentioned that the data available for constructing rating curves in the three catchments were limited to six (three in Xerias) monthly average water levels and their corresponding Q values in January-June 2012 only. These were presumably measured at the same cross section as where the daily data was collected. Also, from personal communication with local sources and observation on location it appears that the river beds are unstable with erosion and sedimentation of up to 30 cm or more at times. Water is furthermore transferred to surface water reservoirs (see the result section) from all three rivers (TEMES S.A., 2009b).
To our knowledge however, Q is measured upstream from these artificial outlet points so they should not have influence over the resultant water balance.
14
Figure 3: The hydrographs of the three rivers with the set hydrological years marked with
crosshatched vertical lines. In black is runoff (R, see below) and in blue the catchment precipitation (P, on the secondary y-axis), both in mm/yr. Notice that the vertical axis for R in Xerias has a larger scale than the other two. For legibility extremes of R have been cut in Sellas (2176 mm/yr the 7 February 2012) and Xerias (several, with the highest being 49 114 mm/yr (!) the 28 December 2009).
0 20 40 60 80 100
0 200 400 600 800 1000 1200
1400 Sellas
0 20 40 60 80 100
0 200 400 600 800 1000 1200
1400 Gianouzaga
P (mm/yr)
0 20 40 60 80 100
0 1000 2000 3000 4000 5000 6000 7000
8000 Xerias
R (mm/yr)
15
For other spatial analyses there was also a digital elevation model (DEM) with a resolution of 30 m and a land use shape file available from previous studies performed by Delrue (unpubl.
data) and Lundholm et al. (2009) respectively (Fig. 2). The catchments were delineated, and their areas estimated, using the DEM and hydrological tools in ESRI ArcGIS. The same software was used for evaluating the land cover distributions in the catchments, as based on the land use shape file. This land use map was made on a regional scale covering the whole of Messinia hence it gives a rough estimation of local distribution only. It should be noted that because the exact locations of the Q gauges were not specified, there might be an offset of these to the delineation points.
Figure 4: (a) Local meteorological station measuring for example T and P and (b) the radar instrument used for measuring river stage. Photo courtesy: TEMES (a) and Victoras Plevrakis (b).
Finally, there was information provided by Konstantine Boulolis, an agronomist active in Gargaliani (a city 7.6 km from Costa Navarino located outside the municipality of Pylos- Nestor, see Fig. 2), on approximate irrigation amounts applied seasonally in the local olive orchards. This estimation was based on his experience and on “irrigation diaries” filled and turned in by his client farmers reporting field characteristics, dates, duration and volumes of irrigation (Fig. 5). Importantly, this is an approximation of applied amounts only.
Figure 5: Example of an irrigation diary filled in by a local farmer and collected by agronomist Konstantine Boulolis in Gargailianoi, Messinia. The first column gives the field ID, the second and third the date and hours of irrigation respectively and the fourth the quantity of water for each tree (translation Magdalini Zampouni). Field characteristics are given in a separate form.
(a) (b)
16
2.3 Methods
2.3.1 The water balances
There is no limit to the possible complexity of water resources modeling. Data demands and uncertainty increase rapidly the more comprehensive the evaluation and the higher the
number of parameters included. The water balance is a straightforward method for evaluating water availability with little data yet sufficient accuracy. It can be used on any temporal and spatial scale, yet again; the finer the resolution the higher the demand for detail and precision.
Depending on the assumptions made, the water balance equation can be set with varying complexity but the basic concept is for inputs to the water system to equal outputs. Assuming no trans-boundary transfer of water, the balance can be written as:
where P is precipitation, ET evapotranspiration, R runoff ( ), ΔS change in storage (including ground water, soil moisture, snow/ice and lakes/rivers/reservoirs) and μ is the balance discrepancy (preferably close to zero) (Senay, et al., 2011). For large spatial scales and long time periods this equation can be reduced into:
while assuming:
(1) The change in storage is zero. This assumption is (most often) valid when studying long time series, at least one year, where temporal variability is balanced (i.e. the system approaches an approximate steady state). The longer the time considered the greater the power of this assumption. When using this form for shorter time periods, such as for individual months, contributions due to ΔS cannot be ignored.
(2) The discrepancy (uncertainty and errors) is negligible and “embedded” in the parameters left.
Evapotranspiration is difficult to measure and is more conveniently evaluated as the residual that closes the balance. Eq. 2 is therefore rearranged into:
which was the simple equation employed in this study on catchment scale and on an annual basis.
17
Ideally the annual water balance would be evaluated for the hydrological year (HY) starting in September when discharge is lowest. However, in order to use the available time period of data optimally, the HY was set to start in May (HYMay). Annual catchment P was estimated through inverse distance weighed averages (IDWA) i.e. through weighing the nearest gauges (Tab. 3) according to their proximity to the catchment center. Thus, the closer the gauge the greater was the influence on the average. Orographic variance caused uncertainty in
interpolation and there was a lack of gauging stations at higher altitudes where P was richer (Tab. 2 and Fig. 6). Moyzaki was the closest equivalent with the highest P and was therefore given additional weight when estimating P in Sellas. For the same reason, i.e. better
representation of the orographic setting, both Navarino and Sgrapa stations were excluded completely from the calculations.
Table 3: The stations (x) used for the IDWA estimations in each catchment.
Navarino Moyzaki Chora Handrinos Sgrapa Kynigos
Sellas - x x x - -
Gianouzagas - - x x - x
Xerias - - - x - x
Figure 6: Annual average precipitation (HYMay, 2009-2012 or the time available) as a function of altitude at the local meteorological stations set up by TEMES.
Navarino
Moyzaki
Chora Handrinos
Sgrapa
Kynigos
500 600 700 800 900 1000 1100 1200
0 100 200 300 400 500
Annual P (mm)
Altitude (m)
18 2.3.2 ET of local vegetation
The ET of local vegetation was required for subsequent land use modeling. For simplicity and relevant accuracy (especially with the crude land use map available), the current land cover
was classified into two categories only, again: olive agriculture (OLV) and “native”
(other) vegetation (NV). These were then characterized by their different ET values, initially taken from literature. All studies reviewed for this purpose (Fig. A1 and marked with asterisks in the reference list) were performed in Mediterranean climate, either in the Mediterranean countries or in California (one in Arizona), U.S. The primary methods employed were i) eddy covariance measurements, ii) different kinds of modeling or iii) water balance approaches. All in all, 21 reported annual numbers were noted and averaged into approximate representative ET values for both OLV and NV.
These literature values were then calibrated to local conditions and scaled to the three catchments. This was possible with the underlining assumption that the relative distribution (represented as a ratio, rlit, Eq. 4) between the literature ET values of regional OLV and NV (ETo_lit and ETn_lit respectively), applies also to the local area. Scaling was done through combining this ratio with the ET balance in each catchment (Eq. 5). This balance states that:
the total volume of ET (ETt*At) from the catchment, which was estimated in the water balance (Eq. 3), equals the summed ET volume from OLV and NV (Ax*ETx). Eq. 4 and 5 were
rearranged and combined into Eq. 6, in which ETn in the local catchments can be
approximated. Based on this result and Eq. 4, ETo was then estimated in a fourth and final step (A = area, t = catchment total, o = OLV and n =NV).
⁄
)
⁄
2.3.3 Hypothetical land use intensification
Focus then fell on possible land use change and on how the introduction of an additional highly evaporative vegetative cover would affect the current land use - and water balances.
This hypothetical land cover (HYP) was assumed to have an ET (ETh) close to PET (scenarios A-C, Tab. 4) resulting from any intensive land use with for example high water demands and/or exhaustive irrigation. The PET was correlated to literature (average of all studies reviewed is 1240 mm/yr) but also to local Priestly-Taylor modeling performed in the Sellas catchment (again 1240 mm/yr) (Klein, unpubl.). Two opposite alternatives of water
management were addressed when introducing HYP: that of maintaining the current
(presumably) sustainable water consumption (Model 1) and that of consuming surface water until no basin outflow remains (Q = 0) (Model 2). Of course there exists a range of plausible water exploitation between these yet focus here is on the end-point scenarios.
Table 4: The three scenarios, A, B and C, of ETh in mm/yr. Scenario C resembles PET, A is set to a lower value close to P (average for all three catchments is 893 mm/yr) and B is their average.
A B C
ETh, mm/yr 900 1075 1250
19 Model 1
Model 1 returned what measures are required in land management regarding the areas of OLV and NV, to compensate for HYP and maintain the current water balance. The principal
underlining base (Eq. 7) of the model was the same as in Eq. 5 yet here also HYP was included (ETh*Ah). Eq. 8 offers a way of substituting for the area of NV (An) through the fact that all sub-areas add up to the total area. It was used also for back-calculating An after retrieving the decreased area of OLV (Ao) in Eq. 9. This equation originates from combining Eq. 7 and 8 and then solving for Ao.
⁄ Model 2
Model 2 addressed the theoretical capacity of accommodating HYP before ET exceeds P if river flow (Q) would instead be completely exhausted for satisfying the HYP water need. It was set up based on the same ET-balance as in Model 1 only this time, the initial land use distribution was considered plus P replaced ETt on the left-hand side since Q was added to total availability. It was run in two sub-sets assuming that i) the HYP replaces all NV first and then OLV (case 1, Eq. 10-11) and ii) only OLV is replaced by the HYP (case 2, Eq. 12-13).
Eq. 11 and 13 originates from solving for Ah_max in Eq. 10 and 12 respectively.
( )
( )
In both models (Eq. 7-13) it was assumed that:
(1) PET, and thus ETh, was the same in all three catchments despite possible difference in their water balances. This should be justified considering the dependence of PET on P and T primarily, which both have low spatial variability.
(2) The ratio between ETo and ETn as derived from literature was the same in all three catchments, which presupposes plants and their physical conditions were similar.
Applying literature values in the first place is a considerable source of uncertainty that will be addressed in later discussion.
(3) ETn was representative of all “other” land uses than olive agriculture.
(4) Ah increased with a percentage of the total and respective catchment area, rather than with a set magnitude equal in all three catchments.
20
(5) The 2009-2012 average P and ET values were representative for the long-term regional conditions.
(6) There was no transboundary exchange of water between the catchments.
2.3.4 Interviews and questionnaires
The qualitative assessment of water resources was performed by reviewing literature, distributing questionnaires and through simple interviews with concerned actors on location (Tab. 5). The purpose was to investigate how local water resources are managed, what the dominating water uses are and what “pressures” exist on water availability - between different actors and in current and future development. The interviews were set up in an open semi- structured way collecting a narrative and leaving open the possibility of adapting questions and discussions to the interviewee. A Dictaphone was brought for the interviews but functional issues unfortunately hindered recordings.
The questionnaire (Appendix B) was put together with the purpose of gathering general thoughts and opinions from local residents and actors. It was made available both in English and Greek (translation to Greek by Katerina Mazi). Tab. 6 lists the occupation, gender, age and residence city of the contributors that filled in the questionnaire. Most have connection to agriculture (two agronomists and 12 farmers), some are active at Costa Navarino and NEO (8 people) and the rest are “other” local stakeholders. Most interviews and questionnaires were collected on location in June-July 2012 but answers were also brought in afterwards via e- mail.
Table 5: Interview dates and the names and occupation/position of the people interviewed.
Date Name Occupation/position
9 July 2012 Efstathia Zontanou Supervising agronomist at the producer’s group Nileas, Chora
10 July 2012 Giorgos Maneas Station manager at NEO from June 2012, former head of the HOS* Gialova lagoon project
28 July 2012 Konstantine Boulolis Agronomist, Gargaliani 10 July 2012 Panagiotis Andrianopoulos
& Nikitas Crikas General and technical managers at DEYA Pylos**
17 July 2012 Vasilis Karakousis Environment and sustainability manager at TEMES
*The Hellenic Ornithological Society, runs environmental protection work at Gialova lagoon among many other things.
** ΔΕΥΑΠ, public-service corporation responsible for drinking water supply and waste water treatment in the municipality of Pylos- Nestor.
21
Table 6: Name, occupation, gender (1=female, 2=male), age and city of residence of the individuals that filled in the questionnaires in the order of collection (CN =Costa Navarino).
Name Occupation Gender Age Residence
1 Konstantine
Bouloulis Agronomist 2 42 Gargaliani
2 Theodoras
Kostadopulos Farmer 2 38 Gargaliani
3 Panagiotis
Panagiotopoulos Farmer 2 54 Floka (ΦΛΟΚΑ)
(Gargaliani)
4 Helen Boulouli Farmer 1 - Gargaliani
5 Ioannis Boulouli Farmer 2 - Gargaliani
6 Ioanna
Karamichalou Buisiness woman 1 47 Gialova
7 - Restaurant owner /
merchant 1 50 Gialova
8 Ioannis Lopas Pharmacist 2 50 Romanos
9 Dimitrios Kajas Agronomist 2 40 Chora
10 - Farmer 2 32 Pyrgos (north)
11 - Farmer / gym teacher 2 44 Chora
12 - Private employee 1 33 Chora
13 Giorgos Melcher Electrical engineer at CN 2 52 Gialova 14 Heleni
Georgiopoulou
Student and farmer’s
daughter 1 20 Chora
15 Dionisis Sampolis Farmer 2 58 Filiatra
16 Nikos Kalivitis Physicist, former NEO
station manager 2 35 (Crete)
17 Giorgos Maneas NEO station manager,
former head of HOS 2 32 Kalamata
18 Dimitrakopoulos
Takis Farmer, Nileas 2 49 Chora
19 - Farmer, Nileas 2 58 Chora
20 - Farmer, Nileas 2 48 Chora
21 - Farmer, Nileas 2 52 Chora
22 - Farmer, Nileas 2 45 Chora
23 Raphaella Tsianti Hotel employee, CN 1 - Pylos
24 - Hotel employee, CN 1 29 Marathopoli
25 - Civil engineer, CN 2 42 Kalamata
26 Georgia (Vlahou) Director of conventions
and events, CN 1 52 Greece
27 Natasa Glaraki Assistant to the Chief
Destination Officer, CN 1 32 Kalamata
22
3. RESULTS
3.1 The water balances
The annual water balance (Eq. 3) for the three considered catchments are shown in Fig. 7 and Tab. 7 below. Because the estimated R in Xerias (averaging 1602 mm/yr) exceeds rainfall input, the water balance cannot be closed and the catchment is therefore not considered in further analysis. In Gianouzagas, time coverage and data quality are clearly limited. However, average water balance terms HYMay 11-12 (HY3) are similar to Sellas, which indicates that available data are consistent.
P data is available throughout the whole period and annual patterns of rainfall are similar across all catchments. Comparing the hydrological years, HY3 is generally the wettest, HYMay
10-11 (HY2) the driest and HYMay 09-10 (HY1) is closer to the average of the three years. The interpolated annual average P across all the catchments is 893 mm and spatial difference (averaging 39 mm) is smaller than the interannual (averaging 97 mm).
Figure 7: The water balance parameters of the three catchments HY1-3: P (black), R (crosshatched) and ET (white). Notice that the vertical axis for Xerias has a larger scale than the other two. All values are in mm.
Table 7: The actual water balance values (P, R (% of P) and ET estimated as P minus R) in the three catchments. All values are in mm/yr.
Sellas Gianouzagas Xerias
HYMay P R ET P R ET P R ET
09-10 891 416 (47) 476 862 - - 915 2148 (235) -
10-11 834 302 (36) 532 839 - - 812 1166 (144) -
11-12 1024 416 (41) 608 928 393 (42) 535 930 1483 (159) - Average 916 378 (41) 539 876 392 (45) 484 886 1602 (181) - 0
200 400 600 800 1000
1200 Sellas
P Q ET
0 200 400 600 800 1000
1200 Gianouzagas
P Q ET
0 400 800 1200 1600
2000 Xerias
P R Amount of water (mm) ET
23
3.2 ET of local vegetation and olive irrigation
Tab. 8 shows a summary of the average ET in OLV and NV obtained through the literature review and the estimated catchment ET values reached after water balance closure and
calibration. As can be seen, OLV has 1.78 times the ET of NV in this type of climate and, even if they differ from the regional estimations, the calibrated catchment numbers are fairly
similar in the two catchments of Sellas and Gianouzagas. ETo and ETn in Gianouzagas amount to 93% of the same values in Sellas while their average is only 80% of the regional
estimation.
Table 8: Regional ET numbers attained from the literature review (ETo_lit and ETn_lit), their ratio (rlit) and the ETo and ETn scaled to the catchments of Sellas and Gianouzagas respectively. All values apart from the ratio (no unit) are in mm/yr.
Regional Sellas Gianouzagas ETo_lit ETn_lit rlit ETo ETn ETo ETn
730 410 1.78 607 341 563 316
In the literature reviewed were also estimations of applied irrigation amounts, all in the Mediterranean region. These range from 181 mm/yr (Fernández, et al., 2006) to 403 mm/yr (Palomo, et al., 2002) and averages 300 mm/yr (additionally: Fernández, et al., 1998; Orgaz and Fereres, 2004; Pastor, 2005; Tognetti, et al., 2006). Also, the average tree density is 223 trees/ha. Local olive fields, according to Mr. Boulolis agronomist in Gargaliani, are irrigated approximately once every second week during the warm and dry season June-September and then harvested in October-December. Generally and based on farmers’ irrigation diaries, the average size of a local farm is 1 ha, the tree density is 200 trees/ha and each tree consumes 1 ton of water per 10 days from July to September. Small trees would need half the water. A rough estimation based on these numbers results in an irrigation amount of around 100-150 mm/yr in the local area (11-17% of total annual rainfall (2009-2012)) depending on land cover distribution, proportion of irrigated agriculture and proportion of “grown” vs. “small”
trees. The extra water evaporated from OLV in relation to NV (estimated above) corresponds to 200 mm/yr and 220 mm/yr on catchment scale in Sellas and Gianouzagas respectively.
3.3 Hypothetical land use intensification
3.3.1 Model 1
The primary result of simulation in Model 1 in Sellas is a clear increase in the ratio of An to Ao
(An/Ao) as HYP covers an increasing proportion of the catchment (Fig. 8, upper chart). This response is similar to that in Gianouzagas (Appendix C and Fig. C1 therein) which supports consistency in the model results. Common for both catchments also is that An/Ao grows more rapidly i) the higher the ETh (compare scenario A to B to C), and ii) the larger the Ah
introduced (growth is non-linear). Also, the larger the scale of development the greater is the influence of ETh.
