Joint GWP CEE/DMCSEE training: From Drought Management Strategies to Drought Management Policies by...

Agricultural University of Athens, Hellas

Christos A. Karavitis

Ass. Professor, Water Resources Management

Sector of Water Resources Management

Dep. of Agricultural Engineering

AUA

From Drought Management Strategies to Drought Management Policies


THE TRANSFORMING ΕNVIRONMENT

TRENDS AND DEVELOPMENTS

PREPARING THE SYSTEM OF THE 21ST CENTURY

SHIFTING PARADIGMS IN THEORY & PRACTICE

SPECULATING ABOUT THE FUTURE

Definition of Drought

A creeping phenomenon, a “non-event”

A source of confusion in devising an objective definition may be that drought implies a variety of things to various professionals according to the specialized field of study (meteorology, hydrology, water resources, agriculture etc.).

• Operational definitions attempt to demarcate the severity, onset and termination point of droughts

• Conceptual definitions attempt to identify the boundaries of the drought event

Drought

• a usually unexpected and unpredicted time period of abnormal dryness which affects water supply" (Grigg, N.S., 1988).

• The state of adverse and wide spread

hydrological, environmental, social and economic impacts due to less than generally anticipated water quantities (Karavitis, 1992)

Social and Economic Drought / Water resources Engineering

• Gap between supply and demand of economic goods such as – water,

– food,

– raw materials,

– hydroelectric power,

– transportation

• depends on the time and space processes of supply and demand

• Social Stresses – Economic impacts

Cook et al., 2004

Percent area affected by drought (PDSI) Western States, USA.

J Sheffield et al. Nature 491, 435-438 (2012) doi:10.1038/nature11575

Global average time series of the PDSI and area in drought. Little change in global drought over the past 60 years

a, PDSI_Th (blue line) and PDSI_PM (red line). b, Area in drought (PDSI <−3.0) for the PDSI_Th (blue line) and PDSI_PM (red line). The shading represents the range derived from uncertainties in precipitation (PDSI_Th and PDSI_PM) and net radiation (PDSI_PM only). Uncertainty in precipitation is estimated by forcing the PDSI_Th and PDSI_PM by four alternative global precipitation data sets. Uncertainty from net radiation is estimated by forcing the PDSI_PM with a hybrid empirical–satellite data set and an empirical estimate. The other near-surface meteorological data are from a hybrid reanalysis–observational data set. The thick lines are the mean values of the different PDSI data sets. The time series are averaged over global land areas excluding Greenland, Antarctica and desert regions with a mean annual precipitation of less than 0.5 mm d−1

The Greenland temperature proxy during the Holocene

Reconstructed from the GISP2 Ice Core (Alley, 2000, 2004). Data from: ftp.ncdc.noaa.gov/pub/data/paleo/icecore/greenland/summit/gisp2/isotopes/gisp2_temp_accum_alley2000.txt

-1.5

-0.5

0.5

1.5

2.5

3.5

01000200030004000500060007000800090001000011000

Years before present

Tem

pera

ture

depart

ure

(oC

)

20-year scale (interpolated)500-year average2000-year average

Medieval

warm period

Little ice age

Roman climate

optimum

Minoan climate

optimum

ANTARCTIC ICE CORE

• Prevailing crisis management attitude

• Natural Hazards Emergency Response Procedures

• Protocols for Processes and Procedures

• Create a wider menu of options and alternatives

Why are Drought Contingency Policies Needed?

Crisis

Management

Risk/Adaptive

Management

NMMC, 2004; Karavitis, C. A., . E. C. Vlachos and N. S. Grigg, 2005

Preparedness

Mitigation

Reconstruction

Recovery Response

Impact Assessment

Disaster

Prediction Warning

Analysis for a Risk Management Framework

Adaptation Planning

Source: IPCC Dr. Thomas E. Downing (SEI)

First Plenary meeting 22-25 November - Nauplion 2011

FP7 Tomorrow’s answers start today

Vulnerability

The degree to which a system, subsystem, or

system component is likely to experience harm

due to exposure to a hazard, either a perturbation

or stress/stressor (Turner et.al, 2003)

Institution Logo

Vulnerability = F(Hazard , Impacts)

UN-ISDR, 2004

The vulnerability term is composed of two basic elements: hazard and impacts. Thus, without a hazard or something to be affected, no vulnerability is implied.

Exposure is also considered to be part of vulnerability though sometimes it is conceived as the relation that connects the system to the concerning hazard.

Generally, vulnerability refers to the factors that affect both the systems likelihood to be harmed and the system’s ability to cope.

Drought vulnerability depends both on the sector of concern where drought is applied on (agriculture, industry, etc.) and the system of interest.

Drought contingency Plan

Primary Meteorological, Agricultural or Hydrological Drought Indices

• Percent of Normal • Deciles • Palmer Drought Index

– PDSI, PHDI, CMI

• Surface Water Supply Index • Standardized Precipitation

Index (SPI) • Reconnaissance Drought Index

(RDI) • Vegetation indices (NDVI, VCI,

SVI) • U.S. Drought Monitor

– Composite index approach

Relation between SDVI & drought aspects (DMCSEE, 2011)

Methodological Framework

• Drought Impact database

• Impact assessment-Resilience

• Vulnerability assessment

• Vulnerability visualization

Steps

1. SPI 6 and SPI 12 calculation and Kriging

2. Supply and Demand Data

3. Impact Data

4. Infrastructure Data on the Level of deficiency

5. Index Calculation

6. Index Visualization

Drought Vulnerability Index

The SDVI is composed of six components in four categories: 1. cSPI-12 representing the non-agricultural water availability (hydropower, households and tourism) and cSPI-6 portraying the agricultural (irrigation) use, respectively. Their values are calculated on local (meteorological station) scale. 2. Supply and Demand that describe the deficits in supplying capacity (network operation and losses) and in demand coverage. Their magnitude depends on the available and delivered amount of water. 3. Impacts that describe the losses (transformed into monetary units) that might have been caused due to the supply – demand deficiencies.


4. Infrastructure that describes the existing infrastructure level of development regarding the level of deficiency (divergence from the designed supply capacity). The values of the components included in the last three categories are calculated on a basin or sub-basin scale (in the case of infrastructure, average values are used). The assessment is based on a synthetic SPI-based Drought Vulnerability Index (SDVI) that was developed by Agricultural University of Athens (2011), in the context of Drought Management Centre in South-eastern Europe (DMCSEE Project-eu).

Components Classification

(DMCSEE, 2011)


The assessment is based on a synthetic SPI-based Drought Vulnerability Index (SDVI) that was developed by Agricultural University of Athens, in the context of Drought Management Centre in South-eastern Europe (DMCSEE Project , 2011). .

𝑆𝐷𝑉𝐼 = 𝑆𝑐𝑎𝑙𝑒𝑑 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝑡ℎ𝑒 𝐶𝑜𝑚𝑝𝑜𝑛𝑒𝑛𝑡𝑠

𝑁

𝑁

𝑖=1

The equation implies that all the components are equally weighted.

Lower Upper Relative Cumulative Cum. Rel.

Class Limit Limit Midpoint Frequency Frequency Frequency Frequency

at or below 0 0 0.0000 0 0.0000

1 0 0.5 0.25 24 0.0563 24 0.0563

2 0.5 1.0 0.75 64 0.1502 88 0.2066

3 1.0 1.5 1.25 189 0.4437 277 0.6502

4 1.5 2.0 1.75 125 0.2934 402 0.9437

5 2.0 2.5 2.25 23 0.0540 425 0.9977

6 2.5 3.0 2.75 1 0.0023 426 1.0000

above 3.0 0 0.0000 426 1.0000

SENSITIVITY ANALYSIS & CLASSIFICATION

Index Classification

Methodology used

7 Case studies were selected located in:

1. Bulgaria

2. F.Y.R.O.M.

3. Hellas

4. Hungary

5. Montenegro

6. Serbia

7. Slovenia

• Water Stress Vulnerability Index (WStVI)

• Water Scarcity Vulnerability Index (WScVI)

• Drought Vulnerability Index (DVI)

For each point (sub-area), the web tool can automatically calculate:

1.3 - Calculation + Visualization

47

Water Changing Conditions

Selection & Classification of Indicators for DVI

Non Vulnerable Less Vulnerable Medium Vulnerable Highly Vulnerable Extremely Vulnerable Exceptionally Vulnerable

0 1 2 3 4 5

Coefficient of variation of precipitation (%) < 20 20 to 30 30.1 to 35 35.1 to 40 40.1 to 50 > 50

Dryness Ratio (%) < 20 20 to 30 30.1 to 40 40.1 to 50 50.1 to 65 > 65

SPI 1 > 0.50 -0.49 to 0.49 -0.50 to -0.99 -1.00 to -1.49 -1.50 to -1.99 < -2.00

SPI 3 > 0.50 -0.49 to 0.49 -0.50 to -0.99 -1.00 to -1.49 -1.50 to -1.99 < -2.00

SPI 6 > 0.50 -0.49 to 0.49 -0.50 to -0.99 -1.00 to -1.49 -1.50 to -1.99 < -2.00

SPI 9 > 0.50 -0.49 to 0.49 -0.50 to -0.99 -1.00 to -1.49 -1.50 to -1.99 < -2.00

SPI 12 > 0.50 -0.49 to 0.49 -0.50 to -0.99 -1.00 to -1.49 -1.50 to -1.99 < -2.00

SPI 24 > 0.50 -0.49 to 0.49 -0.50 to -0.99 -1.00 to -1.49 -1.50 to -1.99 < -2.00

Vegetation cover of basin area (%) > 80 79.9 to 70 69.9 to 60 59.9 to 50 49.9 to 35.0 < 35

Total Rate of Coverage (%) 100 99.9 to 95.00 94.9 to 90 89.9 to 85 84.9 to 80 < 80

Domestic Coverage (%) 100 99.9 to 95.00 94.9 to 90 89.9 to 85 84.9 to 80 < 80

Industrial Coverage (%) 100 99.9 to 95.00 94.9 to 90 89.9 to 85 84.9 to 80 < 80

Agricultural Coverage (%) 100 99.9 to 95.00 94.9 to 90 89.9 to 85 84.9 to 80 < 80

Coverage of Other Uses (%) 100 99.9 to 95.00 94.9 to 90 89.9 to 85 84.9 to 80 < 80

Change in Total Coverage (%) > 3 3 to 0.1 0 -0.1 to -3 -3.1 to -5.0 > -5

Change in Domestic Coverage (%) > 3 3 to 0.1 0 -0.1 to -3 -3.1 to -5.0 > -5

Change in Industrial Coverage (%) > 3 3 to 0.1 0 -0.1 to -3 -3.1 to -5.0 > -5

Change in Agricultural Coverage (%) > 3 3 to 0.1 0 -0.1 to -3 -3.1 to -5.0 > -5

Change in Coverage for Other Uses (%) > 3 3 to 0.1 0 -0.1 to -3 -3.1 to -5.0 > -5

Population Density (No/km2) < 19.1 19.1 to 50.0 50.1 to 90.0 90.1 to 150.0 150.1 to 250.0 > 250

Population Growth (%) < 0 0 to 0.6 0.61 to 1.10 1.11 to 1.60 1.61 to 2.00 >2.00

. . . . . . .

. . . . . . .

. . . . . . .

Drought Vulnerability Indicators (WStVI)

Calculation of Indicators (75 Indicators)

RSPI1 RSPI3 RSPI6 RSPI9 RSPI12 RSPI24 AGRGDP AGRGROW AcSan ChAcSan AgrPr ChAgrPr CosWatExTr . . .

2 1 1 1 1 1 5 1 4 0 0 3 5 .

1 0 2 1 1 1 5 1 4 0 0 3 5 .

0 1 1 1 1 1 5 1 4 0 0 3 5 .

0 1 1 1 1 1 5 1 4 0 0 3 5 .

2 4 2 2 1 2 5 1 4 0 0 3 5 .

2 2 1 1 1 2 5 1 4 0 0 3 5 .

3 1 1 1 1 1 5 1 4 0 0 3 5 .

0 1 1 1 1 1 5 1 4 0 0 3 5 .

1 1 1 1 1 2 5 1 4 0 0 3 5 .

1 1 1 1 0 3 5 1 4 0 0 3 5 .

1 1 1 1 1 3 5 1 4 0 0 3 5 .

0 1 1 1 1 3 5 1 4 0 0 3 5 .

Correlation matrix & Descr. Stats for DVI (52 Indicators)

SPI 1 SPI 3 SPI 6 SPI 9 SPI 12 SPI 24 Population Density (No/km2) Population Growth (%) Agricultural Growth (%) Industrial Productivity ($/m3) Change in Industrial Productivity (%) Agricultural Productivity ($/m3) Total Use of Reclaimed Water (Ratio) Change in Agricultural Productivity (%) Total Stored Water (hm3) Change in Stored Water (%) Gross Domestic Product ($/capita) Water Supplied Population (%) Change in Supplied Population (%) Tourists Accomodation (Days) Tourists Density (No/km2) Population under Poverty (%) Change in Population Under Poverty (%) Water Governance (Q) Wastewater Governance (Q) Degree of Preparedness (Q)

Status of Water Supply Infrastructure (Q) Status of Wastewater Treatment and Supply Infrastructure (Q) Change in Use of Reclaimed Water (%) Legal Framework (Q) Financial Framework (Q) Institutional Framework (Q) Policy Framework (Q) Index of Inflow (%) Storage Capacity per capita (m3/capita) Distribution Network Efficiency (%)

Share of water use by agriculture adjusted by the sector’s share of GDP Share of water use by industry adjusted by the sector’s share of GDP Irrigated land as % of cultivated land (%) Hydropower Dependence (%) Water charges as percentage of per capita income (%) Literacy rate (%) Economically active population (%) Investments (%GDP) Cost of Water Extraction and Treatment ($/m3) Groundwater development stress indicator (GDSI) Share of domestic supply from conventional freshwater sources (surface & groundwater) Share of agricultural supply from conventional freshwater sources (surface & groundwater) Share of industrial supply from conventional freshwater sources (surface & groundwater) Share of domestic supply from alternative freshwater sources (e.g. reclaimed water) Share of agricultural supply from alternative freshwater sources (e.g. reclaimed water) Share of industrial supply from alternative freshwater sources (e.g. reclaimed water) Change in Use of Reclaimed Water (%)

Factor Analysis for Drought Vul. Ind.

Component Initial Eigenvalues Extraction Sums of Squared Loadings

Total % of Variance

Cumulative %

Total % of Variance

Cumulative %

1 3.187 45.522 45.522 3.187 45.522 45.522

2 1.494 21.349 66.871 1.494 21.349 66.871

3 1.091 15.589 82.461 1.091 15.589 82.461

4 .614 8.777 91.238

5 .358 5.119 96.357

6 .185 2.644 99.001

7 .070 .999 100.000

DVI Indicators weights

Index of Flow 0.092

RSPI6 0.166

RSPI12 0.191

Change in Agricultural Productivity (%) 0.078

Change in Industrial Productivity (%) 0.075

Status of Wastewater Treatment and Supply Infrastructure

0.202

Status of Water Supply Infrastructure 0.196

SENSITIVITY ANALYSIS & CLASSIFICATION

Lower Upper Relative Cumulative Cum. Rel.

Class Limit Limit Midpoint Frequency Frequency Frequency Frequency

at or below 0 0 0.0000 0 0.0000

1 0 0.714286 0.357143 64 0.0064 64 0.0064

2 0.714286 1.42857 1.07143 1000 0.1000 1064 0.1064

3 1.42857 2.14286 1.78571 3050 0.3050 4114 0.4114

4 2.14286 2.85714 2.5 3630 0.3630 7744 0.7744

5 2.85714 3.57143 3.21429 1808 0.1808 9552 0.9552

6 3.57143 4.28571 3.92857 412 0.0412 9964 0.9964

7 4.28571 5.0 4.64286 36 0.0036 10000 1.0000

above 5.0 0 0.0000 10000 1.0000

Classes

No Vulnerability 0

Very Low Vulnerability 0.72 1.42 1

Low Vulnerability 1.43 2.14 2

Medium Vulnerability (At Risk) 2.15 2.86 3

High Vulnerability 2.87 3.57 4

Very High Vulnerability 3.58 4.28 5

Extreme Vulnerability 6


<0.71

>4.29

The results for the whole study area can then be represented on the visualization Platform (eg: using Kriging - regression)

1.3 - Calculation + Visualization

54

After this initial analysis, the web tool will allow the end--user to deeply analyze/improve the vulnerable subareas, towards possible WR&R:

2 - WR&R Evaluation and Policy options

55

Nevertheless, the major challenge for any drought related

research may be the development of comprehensive and

effective drought management and decision making schemes. In

such quests, forecasting may provide some help…

• Traditional Planning – Assumes stationary climate

– Uses recorded weather and hydrology times series

– Cylinder of Certainty

Cone of Uncertainty

58 Adapted from Malcolm

Pirnie,2009

Present

Recorded Variability

Future

Updated Planning Methods Hundreds of possible climate scenarios

Multi-outcome planning

Robust over optimal

Scenario Planning

• Small number of equally likely scenarios [A, B, C, D]

• Common strategies (no regrets)

• Sign posts

59

A

B C

D

Sign Post

Present

Future

Robust Decision Making

– CIS analysis of many plausible likely scenarios

– Iteration and hedging

60

Present

Future

DROUGHT FORECASTING

• Statistical predictions: based on lagged relationships between regional

meteorological patterns (i.e. rainfall) and actual data.

• Probabilistic predictions: by trying to estimate the mathematical

probability of a given phenomenon in a certain time frame based usually

on analysis of time series data.

• Deterministic predictions: which attempt to model important features of

the existing system and provide a basis for contingency analysis through

simulation

Greece is highly dependent on the annual rainfall patterns resulting

that any precipitation deficit may cause most of the time significant

impacts on the economy, societal activities and the environment.

A series of such deficits have occurred during the last decades (e.g.

1989-90, 1993, 2000, 2003, 2007) characterizing Greece as drought

prone, exposing economy to threats and leaving it vulnerable to

loses.

Mean Annual Precipitation 1947-2011

ΕΜΙΤ, AUA, 2011

THE ATHENIAN 1989-90 DROUGHT

TECHNICAL SUPPORT TO THE CENTRAL WATER AGENCY FOR THE DEVELOPMENT OF A DROUGHT MASTER FOR GREECE AND AN

IMMEDIATE DROUGHT MITIGATION PLAN

Contract No. 10889/11/07 /2007

Ministry of Planning, Public Works and the Environment

To Water Resources Management Sector,

Department of Natural Resources and Agricultural

Engineering,

Agricultural University of Athens

Co-ordinator Christos A. Karavitis

Submitted on October 2008

the Seasonal ARIMA Model (Auto Regressive

/Integrated/Moving Average) has been applied for

forecasting

Forecasting Methodology

52 stations

24 station at Central Greece for January

2007

Application for all Greece 33 stations

Validation model (Seasonal ARIMA 15 years)

Station X Y Period Average Standard

deviation Minimum Maximum

Stnd.

skewness

Stnd.

kurtosis

Agia Triada 405136 4244800 1981-2010 986.56 220.27 534.3 1382.7 -1.10 -0.30

Amfissa 358886 4265827 1981-2010 681.94 161.66 328.4 1025.7 0.34 0.42

Gravia 363497 4280548 1981-2010 822.31 143.23 431.3 1048.9 -1.87 0.80

Davlia 389166 4248703 1981-2010 770.83 153.20 463.5 1062.1 -0.63 -0.89

Distomo 383406 4253888 1981-2010 605.78 121.66 329.4 821.7 -0.78 -0.45

Eptalofos 367725 4273077 1981-2010 972.64 221.66 612.1 1420.1 0.41 -1.22

Zilefto 349557 4310404 1981-2010 394.19 166.89 96.9 835.7 0.93 0.22

Thisvi 409381 4233654 1981-2010 458.00 149.83 185.6 874.9 0.93 0.90

Itea 363056 4254654 1981-2010 377.45 119.48 187.7 660.9 0.80 -0.53

Kalithea 451708 4238840 1981-2010 468.85 157.18 0 753.6 -1.35 1.61

Kaloskopi 354830 4282551 1981-2010 879.01 235.07 568.0 1740.1 4.01 5.87

K.Tirothea 388071 4274616 1981-2010 655.60 129.32 388.8 991.9 0.31 0.84

Lamia 361050 4306493 1981-2010 538.38 115.08 328.0 870.5 1.12 1.27

Livadia 400880 4254100 1981-2010 737.21 156.38 414.1 1037.8 0.51 -0.43

Lilea 369237 4276752 1981-2010 848.55 335.74 0 1272.4 -3.59 2.44

Pavlos 421355 4264972 1981-2010 479.07 160.67 82.5 741.5 -0.86 -0.23

Trilofo 345367 4317887 1981-2010 598.79 126.42 373.0 826.5 0.72 -0.85

Tymfristos 319174 4309189 1981-2010 1022.99 382.17 141.0 1713.3 -1.55 0.09

Elliniko 475537 4194336 1981-2010 363.63 96.39 155.4 546.6 -0.82 0.06

Asteroskopio 475089 4202597 1981-2010 397.06 138.72 150.6 896.0 3.30 5.36

Elefsina 440916 4212390 1981-2010 344.17 91.06 117.4 468.3 -1.48 -0.16

Tatoi 480783 4218209 1981-2010 435.09 153.46 169.6 812.7 -0.03 -0.15

Filladelphia 477479 4210345 1981-2010 452.17 177.52 180.5 1015.5 2.95 3.35

Ypati 346524 4303061 1981-2010 790.74 293.18 327.6 1431.0 0.98 -0.57

Adopted Methodology

Drought event in 2007… 1. Based on the available raw data obtained from twenty-four meteorological stations

and for a calibration period of thirty years (1981 – 2010), the SPI (3, 6, 9, 12 and 24) values were calculated on a monthly scale.

2. The SPI values from January, 1981 to December, 2006 were used as input data in the ARIMA (Auto Regressive /Integrated/Moving Average) model for a forecasting period of three months (January – March, 2007) to be developed.

3. Comparative analysis of Steps 1 and 2

Divergence of the projected SPI 6 values compared to the observed ones

Application Period (2013-2020)

Station X Y Period Average St. Deviation Coeff. of

variation Minimum Maximum

Stnd.

skewness Stnd. kurtosis

Agrinio 268875.16 4275530.28 1947-2012 891.31 225.80 0.25 388.60 1463.90 0.25 -0.74

Aktion 219283.86 4312709.01 1971-2012 873.47 179.94 0.21 504.40 1251.40 0.21 -0.87

Alexandroupolis 662821.24 4523618.93 1947-2012 540.44 134.46 0.25 275.70 905.90 0.91 -0.60 Anchialos 396082.70 4342228.76 1956-2012 491.37 116.95 0.24 288.30 847.00 2.29 0.61 Andravida 261465.38 4199167.76 1959-2012 778.55 178.68 0.23 325.60 1160.00 -0.73 -0.10 Arta 240224.50 4338878.73 1976-2012 1082.74 232.05 0.21 493.70 1509.90 -0.51 -0.26 Asteroskopio 475089.38 4202597.25 1858-2012 394.97 111.63 0.28 115.70 896.00 5.72 9.20 Chania 512901.68 3932269.85 1959-2012 633.32 171.80 0.27 287.50 977.20 1.12 -1.12 Chios 687028.85 4245693.86 1974-2012 562.24 158.66 0.28 110.30 857.80 -0.98 0.69 Elefsina 440915.88 4212389.83 1951-2012 377.89 111.36 0.29 117.40 740.40 2.00 2.66 Hellenicon 476989.70 4193354.66 1948-2012 367.81 92.76 0.25 155.40 560.30 0.17 -0.30 Heraclion 607395.45 3910363.07 1947-2012 478.71 117.08 0.24 261.80 793.90 0.78 -0.35 Ioannina 227392.02 4398141.29 1950-2012 1096.80 247.13 0.23 490.70 1708.20 0.18 -0.22 Kalamata 324360.97 4102811.24 1971-2012 766.62 167.77 0.22 288.00 1082.30 -1.90 0.82 Kastoria 269220.67 4480918.49 1981-2012 554.86 138.64 0.25 281.47 882.94 0.53 0.11 Kerkyra 147783.91 4391112.67 1947-2012 1119.05 246.07 0.22 620.80 1636.30 0.29 -1.04 Konitsa 256042.84 4462236.91 1955-2012 922.17 304.21 0.33 209.80 1642.00 -0.12 -0.08 Lamia 360383.21 4300968.99 1956-2012 571.76 151.69 0.27 159.28 973.80 0.93 1.69 Larisa 366008.76 4389788.05 1949-2012 431.64 109.36 0.25 211.30 772.50 1.68 1.44 Lesvos 725179.25 4325645.06 1952-2012 642.64 158.24 0.25 246.30 931.30 -0.86 -0.30 Lymnos 605166.73 4419248.05 1968-2012 513.19 175.49 0.34 211.86 970.01 1.32 0.13 Melos 538541.95 4065090.13 1950-2012 407.68 125.18 0.31 69.10 774.80 0.86 1.56 Methone 295151.20 4077583.50 1967-2012 647.01 133.53 0.21 389.66 969.79 0.84 0.03 Rhodes 866517.69 4035914.60 1955-2012 670.42 180.79 0.27 310.70 1112.50 1.11 -0.38 Samos 756986.43 4175249.62 1978-2012 701.20 189.14 0.27 353.00 1064.50 0.08 -1.04 Serres 463452.85 4547812.89 1956-2012 483.42 121.38 0.25 137.30 858.40 1.46 3.10 Skyros 541787.38 4312174.03 1931-2012 471.83 163.61 0.35 159.80 879.50 0.41 -1.13 Tanagra 461951.48 4243016.32 1960-2012 471.21 123.20 0.26 175.60 775.60 0.20 0.10 Tatoi 480783.38 4218208.89 1952-2012 461.30 132.27 0.29 169.60 812.70 -0.18 0.76 Thera 632478.91 4029392.55 1961-2012 324.92 124.81 0.38 16.40 562.90 -0.46 -0.14 Thessalonike 412412.02 4485367.97 1959-2012 454.77 97.97 0.22 248.94 722.87 0.82 -0.09 Tripolis 358423.25 4154541.74 1949-2012 733.92 185.95 0.25 251.98 1192.80 0.04 0.21

Station Model RMSE RUNS RUNM AUTO MEAN VAR

Agrinio ARIMA(2,0,1)x(2,1,2)180 0.101624 OK OK *** OK *

Aktion ARIMA(1,1,2)x(1,2,0)180 0.201477 *** OK

Alexandroupolis ARIMA(2,0,2)x(1,1,2)180 0.186691 OK OK *** ** OK

Anchialos ARIMA(1,1,2)x(2,2,0)180 0.431217 *** OK OK

Andravida ARIMA(1,1,2)x(2,1,0)180 0.135341 *** OK OK

Arta ARIMA(2,0,2)x(2,1,2)180 0.249207 OK * *** OK OK

Asteroskopio ARIMA(2,0,1)x(2,1,2)180 0.343600 OK OK *** OK **

Chania ARIMA(1,1,1)x(2,2,0)180 0.058688 *** OK OK

Chios ARIMA(1,1,2)x(2,0,1)180 0.035345 *** OK OK ***

Elefsina ARIMA(2,0,2)x(2,2,1)180 0.138900 OK *** *** OK

Hellenicon ARIMA(2,0,2)x(2,2,1)180 0.295673 OK *** ***

Heraclion ARIMA(1,1,1)x(1,2,0)180 0.026318 *** OK OK

Ioannina ARIMA(1,0,2)x(2,2,1)180 0.322977 *** ** ***

Kalamata ARIMA(1,1,2)x(2,1,0)180 0.0000510 *** OK OK

Kastoria ARIMA(1,1,1)x(1,1,0)180 0.0014500 * OK OK

Kerkyra ARIMA(2,0,1)x(2,2,2)180 0.3864025 OK OK ***

Konitsa ARIMA(1,1,1)x(2,1,1)180 0.0020666 *** OK ***

Lamia ARIMA(1,1,1)x(2,1,0)180 0.0043118 *** OK OK

Larisa ARIMA(2,0,2)x(2,2,0)180 0.02018869 *** *** ***

Lesvos ARIMA(1,0,1)x(2,2,2)180 0.00440373 *** *** ***

Lymnos ARIMA(1,1,0)x(2,1,0)180 0.0003154 *** OK OK

Melos ARIMA(2,1,1)x(2,2,2)180 0.3947320 OK OK ***

Methone ARIMA(2,0,2)x(2,2,1)180 0.0023547 *** *** ***

Naxos ARIMA(1,1,2)x(2,2,0)180 0.0000037 *** OK OK

Rhodes ARIMA(0,1,0)x(2,2,2)180 0.0001758 *** OK OK

Samos ARIMA(1,1,1)x(1,2,0)180 0.0002198 *** OK OK

Serres ARIMA(1,1,1)x(2,1,0)180 0.0028237 *** OK OK

Skyros ARIMA(2,1,2)x(2,1,2)180 0.4102889 OK OK *** OK **

Tanagra ARIMA(1,1,1)x(2,1,0)180 0.0170188 *** OK

Tatoi ARIMA(2,1,1)x(2,2,2)180 0.0011950 OK OK OK

Thera ARIMA(1,1,1)x(2,2,0)180 0.00054447 *** OK OK

Thessalonike ARIMA(1,2,2)x(2,2,2)180 0.00002195 OK * ***

Tripolis ARIMA(1,1,1)x(2,1,1)180 0.00001420 *** OK OK

RMSE = Root Mean Squared Error RUNS = Test for excessive runs up and down RUNM = Test for excessive runs above and below median AUTO = Box-Pierce test for excessive autocorrelation MEAN = Test for difference in mean 1st half to 2nd half VAR = Test for difference in variance 1st half to 2nd half OK = not significant (p >= 0.05) * = marginally significant (0.01 < p <= 0.05) ** = significant (0.001 < p <= 0.01) *** = highly significant (p <= 0.001)

Residual Autocorrelations for adjusted Col_1

ARIMA(2,1,1)x(2,2,2)180

0 5 10 15 20 25

lag

-1

-0.6

-0.2

0.2

0.6

1

Au

toc

orr

ela

tio

ns

Residual Autocorrelations for adjusted SPI 12

ARIMA(0,1,1)x(2,2,0)180

0 5 10 15 20 25

lag

-1

-0.6

-0.2

0.2

0.6

1

Au

toco

rrela

tio

ns

Residual Normal Probability Plot

ARIMA(0,1,1)x(2,2,0)180

-8 -4 0 4 8(X 1.E-18)

Residual

0.1

1

5

20

50

80

95

99

99.9

pe

rc

en

tag

e

Residual Normal Probability Plot

ARIMA(0,0,0)x(2,2,2)180

-0.1 0.3 0.7 1.1 1.5 1.9 2.3

Residual

0.1

1

5

20

50

80

95

99

99.9

pe

rc

en

tag

e

Will be Drought year the 2015?

NORMAL – WET CONDITIONS IN AUTUMN OF 2017

Conclusions

• Seasonal ARIMA (15 years seasonal length) suggests:

Extreme drought conditions the year 2015

(Cyclades islands, Dodecanese islands & Northern Greece)

Normal – Wet condition the year 2017

• The present effort, as part of other pertinent research, point towards that certain drought

events may be anticipated with some certainty.

• The results produced by ARIMA model may be trusted up to a degree, since forecasting of

complex and random variables seems generally elusive.

• This may be of great value since the forecasting processes can be used for early warning

mechanisms developed as parts of drought contingency planning, monitoring and integrated

management.

• Such mechanisms may be able to minimize the impacts of the various drought events while

increasing the area’s absorbing capacity.

Joint GWP CEE/DMCSEE training: From Drought Management Strategies to Drought Management Policies by...

Environment

Transcript of Joint GWP CEE/DMCSEE training: From Drought Management Strategies to Drought Management Policies by...