Establishing the influence of climate, water extraction and tectonics on the water level 

of the Prespa Lakes (N Greece)

The presented work is part of the project CLIM‐HYDROLAKE (Improving future projectionsof climate change induced hydrological responses by looking into the past: the Lake Prespa /Aliakmonas River case study in Greece). This project is supported by the European Communityunder a Marie Curie Career Integration Grant (Framework Program 7, Grant 321979)

Tim van der Schriek & Christos GiannakopoulosNational Observatory of Athens

Institute for Environmental Research and Sustainable Development

In collaboration with  

Why is the water level of Lake Prespa falling?

Caused by water extraction, groundwater outflow changes and/or climate? 

Threat to: global biodiversity & regional water resources (Lakes Prespa and Ohrid; Drim River and Lake Skardar)

Urgent need to understand Lake Behaviour – future extraction, climate change impacts?

842,00

844,00

846,00

848,00

850,00

852,00

854,00

Ιαν-51

Ιαν-56

Ιαν-61

Ιαν-66

Ιαν-71

Ιαν-76

Ιαν-81

Ιαν-86

Ιαν-91

Ιαν-96

Ιαν-01

Ιαν-06

Ιαν-11Stag

e he

ight

(m.a

.s.l.

FYR

OM

)

Years

Monthly Level of Lake Megali Prespa

Methods & Data

Aim: to explain lake level fluctuations with available data

Documentary analyses, basic statistics, regression analyses, spreadsheet‐based analyses of extraction impact

Most comprehensive analyses so far,covering the entire catchment (SPP data base) for 1951‐2004

• Rainfall  (7 stations; at lake level) • Water extraction data (compilation)• Evaporation (open‐pan; extended by “Penman” 

formula)• Lake level Megali Prespa (FYROM data)• Fluvial Discharge  (Brajcinska River  only)

• Estimates for groundwater flow• Earthquake measurements• Snow records

Lake Water Balance & Geology

Main elements of the Lake Water BalanceIn: Direct Precipitation, Fluvial Discharge, GroundwaterOut: Evaporation, Extraction and Karst Drainage

Geology / Tectonics: Internally draining basin, surrounded by mountains (2400m)All groundwater  / fluvial discharge is locally generated

Intrusive rocks (N,E,basement) – aquicludesLimestone overlying in W (aquifer), downfaulted SW (lake)Subsiding Basin, alluvium (N,E) with small aquifers

Underground outflow only in SW area of lake (Ohrid)Mikri Prespa is separated by an alluvial isthmusEarthquakes  (3‐4 Richter) frequent

Direct precipitation

Evaporation

Fluvial discharge

Groundwater

Karst drainage

Extraction

Rainfall, snowfall and evaporation

‐200

‐100

0

100

200

300

400

1 2 3 4 5 6 7 8 9 10 11 12

Min

Avg

Max

0

20

40

60

80

100

120

1 2 3 4 5 6 7 8 9 10 11 12

0

50

100

150

200

250

1 2 3 4 5 6 7 8 9 10 11 12

Min

Avg

Max

Average monthly rainfall at Lake Level

Monthly evaporation at Lake Level

0

20

40

60

80

100

120

1 2 3 4 5 6 7 8 9 10 11 12

Min

Avg

Max

Monthly snowfall (Vrondero)Monthly rainfall‐evaporation balance at Lake Level

Mediterranean climate with continental influences

There is a positive rainfall‐evaporation balance from Oct‐Apr (c. 860m ‐ near Lake Level) – “wet season”

These months also receive significant snowfall (at Vrondero; 1000m)

Rainfall / snowfall must be much higher in the surrounding mountains (up to 2400m)

Hydrological characteristics

-600-400-2000

200400600800

1 5 9 13 17 21 25 29 33 37 41 45 49 53

Diff

eren

ce (m

m)

Hydrological Years (Oct 1951 – Sept 2004)

Annual Rainfall minus Evaporation Balance1951‐2004: annual rainfall‐evaporation balance at lake level is mostly negative

Suggests significant fluvial input (and minor input from direct snow fall?)

0

1

2

3

4

5

6

1 2 3 4 5 6 7 8 9 10 11 12

Min

Avg

Max

Monthly Runoff Data [m³/s], Brajcinska River, FYR of Macedonia, 1961‐2004

Apr‐May are the months of max. discharge – fed by snow‐melt.

Max. monthly rainfall: Nov‐DecMax. monthly snowfall: Dec‐Jan

‐0,2

‐0,15

‐0,1

‐0,05

0

0,05

0,1

0,15

1 2 3 4 5 6 7 8 9 10 11 12

Average monthly water level fluctuations (cm): Lake Megali Prespa (1951‐2011)

Volumetric conversion

Lake level change (m) needs to be converted to volumetric change (m3) to compare different years – thus correcting for the difference in bathymetry at different stage heights.

842,00

844,00

846,00

848,00

850,00

852,00

854,00

Ιαν-51

Ιαν-56

Ιαν-61

Ιαν-66

Ιαν-71

Ιαν-76

Ιαν-81

Ιαν-86

Ιαν-91

Ιαν-96

Ιαν-01

Ιαν-06

Ιαν-11

Stag

e he

ight

(m.a

.s.l.

FYR

OM

)

Years

Monthly Level of Lake Megali Prespa

This conversion is based on the DEM of the basin;  the DEM for the upper part of the lake is most reliable – and thus annual volumetric differences have been calculated

Water level fluctuations are a function of in‐/outflow factors AND bathymetry

Volumetric changes

Lake volumetric change  per hydrological year (October year 1 to October year 2) – as is common for Mediterranean river/lake systems and ones dominated by snow‐melt

Highly variable, perhaps increase in duration/intensity of large negative volumetric changes

‐600,00

‐400,00

‐200,00

0,00

200,00

400,00

600,00

800,00

1000,00

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59

Annual volumetric change (106 m3) of Lake Megali Prespa (1951‐2011)

Annual Extraction

Extraction data (incl. groundwater abstraction) have been collected from various sources. A time‐series of abstraction was reconstructed for 1951‐2004 (best estimate). Extraction prior to 1979: 20‐30 103m3, after 1979: 10‐14 103m3

‐600,00

‐400,00

‐200,00

0,00

200,00

400,00

600,00

800,00

1000,00

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53

Extraction after 1979: max ~0.005% of total lake volume

Lake volumetric change without extraction (observed annual volumetric change + extracted volume)

Observed lake volumetric change

Extraction: Impact on Lake Level

Reconstructed lake level change (no extraction scenario). Annual extraction was added to annual lake volumetric change (both in m3); the resulting value was converted to Lake level change in m. Subsequently the lake surface area difference was calculated between actual/computed lake level; annual evaporation over the extra surface area was calculated and subtracted from the computed lake level. 

Extraction has a significant cumulative impact on lake level (not visible when looking at annual volumetric changes)

Balance is reached after multiple decades (lake surface evaporation decrease due to lake surface area shrinkage equivalent to extracted volume)

838

840

842

844

846

848

850

852

854

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53

Stage Height (m.a.s.l.)

Annual lake level 1951‐2004 (Oct)

Reconstructed (no extraction and extra evaporation; top) vs observed (bottom) hydro‐yearly stage height change (Oct) of Lake Megali Prespa (1951‐2004).

Theoretical calculations (Lake Megali Prespa) support this analysis. 

Average lake level (no extraction): ~850.6m from 1951 to 1987. This corresponds to a lake surface area of 277.8 km2. Annual evaporation over this surface area: 0.832m *277.8 km2 = 231.1 106m3. Subtracting 14 106m3 (yearly lake water extraction) from 231.1 106m3 gives 217.1 106m3 for the required lake surface evaporation to stabilise lake level. Dividing 217.1 106m3 by 0.832m gives a lake surface area of 260.9 km2 at a lake stage height of ~845.4m. 

Average lake level (no extraction): ~846.5m after 1995. This corresponds to a lake surface area of 266.1 km2; annual evaporation over this surface area is 0.832m *266.1 km2 = 221.4 106m3. If we subtract 14 106m3 (yearly lake water extraction) from 221.4 106m3 we obtain a value of 207.4 106m3 for evaporation to stabilise lake level. Dividing 207.4 106m3 by 0.832m gives a surface area of 249.3 km2 at a lake stage height of ~844.5m. 

Note that lake level falls are dependent on the bathymetry!

Groundwater & Tectonics

Groundwater flows poorly known: inflow is estimated at 10‐60 106m3 per year. The most reliable outflow estimates are 335‐385 106m3 per year. Recent modelling suggests steady outflow rates, except for : Oct 1978 – Sep 1986 (8 years: 606 106m3 total reduction in outflow, or extra inflow), and Oct 1989 – Sep 1996 (7 years: 505 106m3 total reduction in outflow). Extra inflow following fall?

The timing of significant lake level changes does not coincide with recorded earthquakes. There are no R>4 earthquakes; R3‐4 earthquakes occur regularly throughout the observation period (not concentrated during periods of significant lake level change).

842,00

844,00

846,00

848,00

850,00

852,00

854,00

Ιαν-51

Ιαν-56

Ιαν-61

Ιαν-66

Ιαν-71

Ιαν-76

Ιαν-81

Ιαν-86

Ιαν-91

Ιαν-96

Ιαν-01

Ιαν-06

Ιαν-11St

age

heig

ht (m

.a.s

.l. F

YRO

M)

Years

Monthly Level of Lake Megali Prespa

Relationship Lake Level ‐ Climate

y = 0,6708x + 335,84R² = 0,8709

0,0

200,0

400,0

600,0

800,0

1000,0

1200,0

‐600,0 ‐400,0 ‐200,0 0,0 200,0 400,0 600,0 800,0 1000,0

Oct‐Apr R‐E (106m3 )

Hydro‐yearly Lake Volumetric Change without Extraction (106m3)

Lake Volumetric Change vs R‐E (1951‐2004)

y = 0,6708x + 349,91R² = 0,9033

0,0

200,0

400,0

600,0

800,0

1000,0

1200,0

‐600,0 ‐400,0 ‐200,0 0,0 200,0 400,0 600,0 800,0 1000,0

Oct‐Apr R‐E (106m3 )

Hydro‐yearly Volume Change without Extraction & with Inflow (106m3)

Lake Volumetric Change vs R‐E (1951‐2004)

Annual lake level change (Oct‐Oct) isstrongly related to the cumulativeOct‐Apr rainfall‐ evaporation balance

Correlations improve when corrections for extraction and extra inflow are incorporated

If more months are included correlations become poorer

Points representing lake level falls are close to the regression line (thus not supporting a tectonic cause)

Regression analysis between annual lake volumetric change and Brajcinska River Discharge has an R2 of 0.74 (catchment: 10% of Prespa). Snow‐melt fed discharge exerts a strong control on lake level.

Discussion

Falling lake levels lead to: i) the reduction of lake volume ‐ thus increasing pollutantconcentrations (strongly accelerating the current eutrophication), and ii) reducedunderground outflow to lake Ohrid. The entire Prespa‐Ohrid‐Drim catchment may beaffected with significant consequences for water resources and biodiversity.

The SE Balkans experience decreasing snowfall and an increase in extreme(drought) events; future climate change scenarios suggest this trend to continue.

Lake level variations are therefore expected to increase while lake levels will follow adownward trajectory (that is strongly influenced by the bathymetry.).

Even minor water extraction has a progressive and serious impact on lake level, with a lag‐time of multiple decades (dependent on the bathymetry). Extraction is responsible for 30%of the overall fall since 1987.

It is likely that water extraction will increase under future climate change scenarios, thus amplifyingclimate change impacts on lake level.

The 1962‐1963 extreme lake level rise matches a wet event that is recognisedthroughout Greece; significant falls (1974‐1978 & 1987‐1995) correspond with majorMediterranean‐wide droughts. Levels of e.g. Lakes Skardar, Ohrid & Dorjan also fall.

This regional synchronisation, suggests that local tectonics do not control the majorPrespa lake level fluctuations ‐ [1] the lack of correlation between earthquakes‐lake level changeand [2] the good correlation of lake volume change with precipitation support this conclusion .

Conclusions

The presented linear regression and extraction impact models will helpsteer adaptation and mitigation strategies by informing on lake responseunder different climate change and extraction scenarios

There are two key issues to should be the focus of any strategies dealingwith climate change impacts on lakes in the S Balkans: address the causesof pollution / eutrophication (to protect water quality and biodiversity)and decrease water extraction through better irrigation techniques (toensure future water availability and biodiversity)

• This work proves for the first time that annual fluctuations in Prespa Lake Levelare strongly related to the cumulative precipitation during the first 7 months ofthe hydrological year (Oct‐Apr).

• The long‐term falling lake level trend is amplified by minor water extraction(which is not visible from regression analyses of annual data)

• Abrupt, large falls in lake level may induce one‐off emptying of groundwaterstores (through the adjustment of the groundwater table in the adjacent aquifers)

• There is no link between earthquake‐occurrence over the observation period(1951‐2004) and major lake level fluctuations.

THANK YOU

Any Questions?