irrigation management and solutions to cope with agricultural drought in bulgaria
DESCRIPTION
Abstract Ref. № 1452414 Paper № NH21B-1593. Z . Popova * , М. Ivanova*, K . Doneva * , M. Kercheva *, L . S . Pereira**, V . Alexandrov***, P . Alexandrova *. - PowerPoint PPT PresentationTRANSCRIPT
IRRIGATION MANAGEMENT AND SOLUTIONS TO COPE WITH AGRICULTURAL DROUGHT IN BULGARIA Z. Popova*, М. Ivanova*, K. Doneva*, M. Kercheva*, L.S.Pereira**, V.Alexandrov***, P. Alexandrova*
AGU Fall Meeting 3-7 December 2012, San Francisco, California
*N. Poushkarov Institute of Soil Science, Agrotechnology and Plant Protection-ISSAPPNP, 7 Shosse Bankya Str., 1080 Sofia, Bulgaria ([email protected]); **CEER-Biosystems Engineering, Institute of Agronomy, Technical University of Lisbon, Tapada de Ajuda, 1349-017 Lisboa, Portugal
*** National Institute of Meteorology and Hydrology- NIMH, 66 Tsarigradsko chaussee Blvd., 1784 SofiaThe objective of this study is to quantify the impact of climate variability, including drought, on agricultural productivity and irrigation requirements in Bulgaria using simulation model WinISAREG for maize crop and seasonal SPI2 “July-Aug” for the period 1951-2004.
AcknowledgementsWe gratefully acknowledge the financial support of Drought Management Center for South East Europe Project, Contract no. SEE/A/091/2.2/X, South East Europe Transnational Cooperation Programme co-funded by the European Union, for implementation and dissemination of our studies’ results on crop vulnerability to droughts and irrigation management in Bulgaria.
Materials and methods:
Climate
-3
-2
-1
0
1
2
3
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004a)
-3
-2
-1
0
1
2
3
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004b)
Fig.1 .Evolution of High Peak Season (July-Aug) SPI2 at: a) Sofia and b) Plovdiv, 1951-2004.
Soil
0
20
40
60
80
100
120
140
May June July August September
mo
nth
su
m o
f p
reci
pit
atio
n (
mm
)
a)0
20
40
60
80
100
120
140
May June July August September
mo
nth
su
m o
f p
reci
pit
atio
n (
mm
)
b)
CropMaize was selected as a typical summer crop. Crop coefficients Kc, water depletion fraction for no stress p and the yield response factor Ky (Allen et al., 1998) were calibrated and validated using detailed independent datasets relative to long term experiments with late and semi - early maize varieties carried out under different irrigation schedules in Tsalapitsa, Plovdiv, Pustren and Zora, Stara Zagora, and Bojurishte, Sofia field.
Simulation model
The model was previously validated for maize hybrids of different sensitivity to water stress on soils of small, medium and large total available water (TAW) in various locations of Bulgaria (Popova et al., 2006; Popova, 2008; Popova and Pereira, 2010; Ivanova and Popova, 2011).
Fig.3 Net irrigation requirements (NIRs) probability of exceedance curves relative to soil groups of small, medium and large total available water (TAW) at: a) Sofia field; b) Plovdiv, South
Bulgaria, maize, 1951-2004.
a)
Fig.8 Relationships between seasonal SPI2 “July-Aug” (X-axis) and relative yield decrease of rainfed maize RYD with Ky=1.6 (Y-axis) at: a) Plovdiv and b)
Pleven; soils of large TAW (180 mm m-1), late maize hybrids.
Fig.9 Map of economical SPI2 “July-Aug” threshold, under which soil moisture deficit leads to severe impacts on yield losses for rainfed maize, Bulgaria.
Results from Table 3, Soil map and Map of TAW of the soils are used for mapping the SPI2 “July-Aug” threshold.
0
10
20
30
40
50
60
70
80
90
100
1993
2000
1952
1962
1965
1988
2001
1987
1992
1958
1994
1961
1974
1985
1956
2003
1990
1982
1954
1999
1963
2004
1969
1970
1997
1973
1953
1964
1966
1996
1978
1981
1979
1975
1998
1959
1984
1986
1983
1989
1957
1967
1955
1980
1960
1971
1972
1977
1968
1991
2002
1995
1976
1,43 5 7 9111315161820222426283031333537394143454648505254565860626365676971737577788082848688909293959799
Year
PRYD (%)
RY
D (
%)
49% risky years (26/53) for semi early hybrid (Ky=1.6)
0
10
20
30
40
50
60
70
80
90
100
2000
1993
1965
1952
1994
1987
1996
2003
1992
1990
1988
1962
1954
1985
1964
1981
1982
1998
1973
1958
1953
1995
1968
1999
1956
1986
1980
2001
1970
1969
1963
1997
1978
1984
1974
1972
1989
1967
1951
2004
1966
1960
1979
1991
1957
1975
1983
2002
1977
1976
1961
1955
1971
1959
1,43 5 7 911121416182022242527293133353638404244464749515355575960626466687071737577798182848688909294959799
Year
PRYD (%)
RY
D (
%)
71% risky years (38/54) for the late hybrids (Ky=1.6)
a)
b)
0
10
20
30
40
50
60
70
80
90
100
1993
2000
1952
1962
1965
1988
2001
1987
1992
1958
1994
1961
1974
1985
1956
2003
1990
1982
1954
1999
1963
2004
1969
1970
1997
1973
1953
1964
1966
1996
1978
1981
1979
1975
1998
1959
1984
1986
1983
1989
1957
1967
1955
1980
1960
1971
1972
1977
1968
1991
2002
1995
1976
1,4 3 5 7 9 111315161820222426283031333537394143454648505254565860626365676971737577788082848688909293959799
Year
PRYD (%)
RYD Threshold Observed RYD H708 (Vurlev, Kolev, Kirkova)Simulated RYD with Ky=1.6 Observed RYD H708 (Rafailov)Observed RYD for semi-early hybrids, (Jivkov and Vurlev)
49% risky years (26/53) for semi early hybrid with Ky=1.6
Fig.5 Probability of exceedance curve of relative yield decrease RYD for rainfed maize, Ky=1.6, soil of small TAW (116 mm m-1) at: a) Chelopechene, Sofia field, and b) Tsalapitsa, Plovdiv, 1951-2004.
0
10
20
30
40
50
60
70
80
90
100
2000
1994
1993
1965
1988
1996
1962
1981
1992
1958
1995
1980
1978
1972
1984
1963
1969
1982
1967
1951
2004
1979
1960
1997
1977
1976
1961
1955
1971
2002
1959
1983
1,43 5 7 911121416182022242527293133353638404244464749515355575960626466687071737577798182848688909294959799
Year PRYD (%)
RY
D (%
)
65 % risky years (35/54) for TAW=11657 % risky years(31/54) for TAW=13632 % risky years(17/54) for TAW=180
a)
0
10
20
30
40
50
60
70
80
90
100
1958
2000
1993
1963
1965
2003
1988
1952
1961
1990
1994
1996
1962
1974
1980
1956
1989
1986
1964
1951
1997
1972
1981
1968
1960
1978
1971
1983
1995
1955
1991
1979
1970
1977
1957
1975
2002
1,43 5 7 911121416182022242527293133353638404244464749515355575960626466687071737577798182848688909294959799
Year PRYD (%)
RY
D (%
)
32 % risky years (17/54) for TAW=11619 % risky years(10/54) for TAW=136-157
13 % risky years(7/54) for TAW=180
a) b)
0
10
20
30
40
50
60
70
80
90
100
1958
2000
1993
1963
1965
2003
1988
1952
1961
1990
1994
1996
1962
1974
1980
1956
1989
1986
1964
1951
1997
1972
1981
1968
1960
1978
1971
1983
1995
1955
1991
1979
1970
1977
1957
1975
2002
1,43 5 7 911121416182022242527293133353638404244464749515355575960626466687071737577798182848688909294959799
RYD Threshold late hybrids TAW=180 mm m-1 TAW=136-157 mm m-1 TAW=116 mm m-1
32 % risky years (17/54) for TAW=11619 % risky years(10/54) for TAW=136-157
13 % risky years(7/54) for TAW=180
Fig.6 Probability exceedance curves of RYD of rainfed maize for soil groups of small, medium and large TAW, Ky=1.6, at: a) Plovdiv, South and b) Pleven, North Bulgaria, late
maize hybrids (H708, 2L602, BC622), 1951-2004.
b)
a)
Fig.7 Comparison of relative yield decrease (RYD, %) probability of exceedance curves, Ky = 1.6, six climate regions and two soil groups of: a) medium (136-157
mm m-1) and b) large (180 mm m-1) TAW, rainfed maize, 1951-2004.
South Bulgaria North Bulgaria
Sofia field Thracian Lowland Sandanski Danube Plain
Sofia Plovdiv Stara Zagora Sandanski Pleven Lom Silistra Varna Climate region
Continental
Transitional continental
Transitional Mediterranean Continental
Black sea
Maize hybrid semi early Late (H708, 2L602, BC622)
TAW mm m-1
Average Yield, kg ha-1
Cv, %
Average Yield, kg ha-1
Cv, %
Average Yield, kg ha-1
Cv, %
Average Yield, kg ha-1
Cv, %
Cv, %
Cv, %
Cv, %
Cv, %
116 4421 42 3894 69 3723 59 2292 72 50 55 46 50 136-157 4920 37 4550 59 4299 52 2906 59 44 47 40 42
180 5896 29 5915 43 4250 41 34 35 30 30 173 5483 41
Table 1. Variability of rainfed maize grain yield characterized by the average value, kg ha-1, and the coefficient of variation Cv, %, climate regions and soil groups in Bulgaria, 1951-2004.
Table 2. Parameters of specific relationships y=a+bx between simulated relative yield decrease RYD (%) / net irrigation requirements NIRs (mm) and High Peak Season SPI2 “July-Aug” across considered soil groups and climate regions, Bulgaria, maize, 1951-2004
Table 3. Economical threshold of High Peak Season SPI2 “July-Aug” (the average SPI2 for July and August) indicating the risk relative to rainfed maize for climate region/soil groups in Bulgaria
0
20
40
60
80
100
120
140
May June July August September
wet year with P=10% average year with P=50% dry year with P=90%
Conclusions:
Rainfed maize yield and risky years
Deriving of drought vulnerability categories
Drought vulnerability mapping
The study relative to eight climate regions, three soil groups and the period 1951-2004, Bulgaria, shows that:
٠In soils of large TAW (180 mm m-1), Plovdiv (Transitional Continental climate), net irrigation requirements (NIRs) range 0-40 mm in wet years and 350-380 mm in dry years. In soils of small TAW (116 mm m-1), NIRs reach 440 mm in the very dry year (Fig.3b). NIRs in Sofia and Silistra (Continental climate) are about 100 mm smaller than in Plovdiv while in Sandanski and Northern Greece (Transitional Mediterranean climate) they are 30-110 mm larger (Fig.4). ٠Rainfed maize is associated with great yield variability in this country (29%<Cv<72%). The smallest Cv refers to Sofia field for soils of large TAW (29%) while Cv=42% is typical for soils of small TAW there. The most variable yields are found in Sandanski (Cv=72%) and Plovdiv (Cv=69%) if TAW=116 mm m-1. The variability of rainfed maize yield in the Danube Plain (30<Cv<55%) is much lower than in the Thracian Lowland (Table 1).
٠In Plovdiv region reliable relationships (R2>91%) were found for seasonal agricultural drought relating the SPI2 for “July-Aug” with the simulated RYD of rainfed maize (Fig.8) while in Stara Zagora, Sandanski and Sofia the relationships were less accurate (73<R2<83%, Table 2). The study found statistically significant correlations between SPI2 “July-Aug” and simulated RYD of rainfed maize for North Bulgaria (R2>0.81) as well.
٠When maize is grown without irrigation on soils of large TAW maize development is less affected by the water stress and economical losses are produced if high peak season SPI2 is less than +0.20 in Sandanski, -0.50 in Plovdiv and Stara Zagora and -0.90 in Sofia field (Table 3). This threshold ranges between -0.75 (Lom) and -1.50 (Pleven) for North Bulgaria. Corresponding NIR thresholds were identified.
٠The derived reliable relationships and specific thresholds of seasonal SPI2 “July-Aug”, under which soil moisture deficit leads to severe impact of drought on rainfed maize yield for the main climate regions and soil groups in Bulgaria, are representative of a wider area of Moderate Continental, Transitional Continental/Transitional Mediterranean and Black Sea climate in SEE. They are used for elaboration of drought vulnerability maps and identification of drought prone territories at regional and national level (Figs.9 and 10).
SPI2 “July-Aug”
NIR
Irrigation requirements
Fig.4 Comparison of Net irrigation requirements (NIRs) probability curves relative to six climate regions and soil group of medium TAW, 136-157 mm m-1, maize, 1951-2004.
0
50
100
150
200
250
300
350
400
450
500
0 10 20 30 40 50 60 70 80 90 100 PNIRs (%)
NIR
s (m
m)
Pleven Silistra Sofia Plovdiv Sandanski Varna
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
PRYD (%)
RY
D (%
)
12 % risky years (6/51) for Pleven 10 % risky years (5/51) for Silistra 20 % risky years(10/51) for Sofia
29 % risky years (15/51) for Plovdiv 63 % risky years(32/51) for Sandanski
14 % risky years(7/51) for Varna0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
Pleven RYD Threshold Plovdiv/Silistra RYD Threshold Sofia RYD Threshold Pleven Silistra Sofia Plovdiv Sandanski Varna
12 % risky years (6/51) for Pleven 10 % risky years (5/51) for Silistra 20 % risky years(10/51) for Sofia
29 % risky years (15/51) for Plovdiv 63 % risky years(32/51) for Sandanski
14 % risky years(7/51) for Varna
Popova Z.(Ed.) 2012 RISK ASSESSMENT OF DROUGHT IN AGRICULTURE AND IRRIGATION MANAGEMENT THROUGH SIMULATION MODELS
Abstract Ref. № 1452414
Paper № NH21B-1593
Soil groups according to TAW
Small Medium Large Climate Region 116 mm m-1 136-157 mm m-1 173-180 mm m-1
RYD
% NIRs mm
RYD %
NIRs mm
RYD %
NIRs mm
Sandanski
Intercept a 79.6 312,3 74.1 294.2 62.1 256.6 Slope coefficient b -15.0 -66,4 -15.7 -66.2 -16.1 -65.8 R2 (%) 75 70 77 70 78 70 Stara Zagora Intercept a 67.0 259.2 61.81 243.3 51.3 211.4 Slope coefficient b -19.9 -83.2 -20.5 -83.1 -20.5 -83.4 R2 (%) 80 77 82 78 83 79 Plovdiv Intercept a 65.2 244.4 59.4 226.7 47.2 189.9 Slope coefficient b -24.8 -97.3 -24.7 -96.4 -23.6 -93.8 R2 (%) 92 89 92 89 91 89 Lom Intercept a 57.7 202.5 51.3 184.7 38.5 148.7 Slope coefficient b -23.8 -81.7 -23.6 -81.5 -22.1 -78.9 R2 (%) 86 80 86 80 86 81 Sofia Intercept a 48.4 178.8 42.6 162.5 31.2 129.2 Slope coefficient b -21.0 -78.2 -20.5 -77.1 -18.8 -73.3 R2 (%) 76 76 75 75 73 73 Silistra Intercept a 56.1 190.3 49.1 171.7 35.9 135.4 Slope coefficient b -20.5 -68.5 -20.3 -68.5 -19.3 -67.4 R2 (%) 86 84 86 85 86 85 Pleven Intercept a 53.5 202.4 47.6 184.8 35.7 148.4 Slope coefficient b -23.3 -88.1 -23 -87.1 -21.6 -84.3 R2 (%) 82 77 81 76 79 75 Varna Intercept a 63.6 212.3 56.9 195 43 158.1 Slope coefficient b -18.1 -59.6 -17.6 -58.7 -16.5 -56.72 R2 (%) 82 73 81 74 80 73
b)
y = -23.61x + 47.22
R2 = 0.91
0
60
-2,00 -1,50 -1,00 -0,50 0,00 0,50 1,00 1,50 2,00 2,50
SPI (2) for "July - Aug"
RY
D r
aim
fed
mai
ze (
%)
RYD threshold
y = -21.57x + 35.72R2 = 0.79
0
67
-2,50 -2,00 -1,50 -1,00 -0,50 0,00 0,50 1,00 1,50 2,00 2,50
SPI (2) for "July - Aug"
RY
D r
aim
fed
mai
ze (
%) RYD threshold
b)
b)
a) b)
A version of seasonal standard precipitation index SPI (Pereira et al., 2010), that is an average of the index during periods of crop sensitivity to water stress, is used as a specific drought indicator. Average SPI2 for several periods referring to maize vegetation season “May-Aug”, Peak Season “June-Aug”, and High Peak Season “July-Aug” were used to define categories of agricultural drought.
The usual soils in South Bulgaria are the chromic luvisols/cambisols of medium total available water TAW (136 mm m-1) and the vertisols of large TAW (170≤TAW≤180 mm m-1). The typical soils in the plains of North Bulgaria are the chernozems of medium to large TAW, 157-180 mm m-1, and the vertisols (TAW≥170 mm m-1). Alluvial/deluvial meadow and light-textured luvisol soils of small TAW≤116 mm m-1 are well identified over the terraces along the rivers.
Climate uncertainty affects performance and management of agriculture. Extreme weather events, as drought, lead to substantial increase in agricultural risk and unstable farm incomes. The necessity to develop methodologies and simulation tools for better analysing/forecasting/managing the risk of agricultural drought is evident after the extremely dry 2000, 2007 and 2012.
Simulations are performed under a Transitional Continental (Plovdiv and Stara Zagora), a Moderate Continental (Sofia, Pleven, Silistra and Lom), a Transitional Mediterranean (Sandanski) and a North Black Sea (Varna) climate.
Specific drought indicators, relationships and economical thresholds are derived by relating seasonal SPI-index to simulation results of water stress impacts on rainfed maize yield and irrigation requirements (Fig.8, Tables 2 and 3).
Probability curves of maize net irrigation requirement (NIRs, mm) were built for each climate region and soil group using ISAREG model simulations over the period 1951-2004 (Figs. 3 and 4).
Relative yield decrease (RYD, %) is simulated with option ‘maize without irrigation’ and yield response factor Ky = 1.6 for a soil of small TAW (116 mm m-1), Sofia and Plovdiv (Fig.5). Additional RYD data from long-term field experiments carried out with semi-early maize hybrids (Fig.5a) or late maize hybrids H708 (Fig.5b) are plotted as well. Results show that the adopted Ky=1.6 could reflect well the impact of water stress on rainfed maize yield in this country.
When soil water holding capacity ranges (116<TAW<180 mm m-1) the relative yield decrease differs by about 20% (Fig.6).
Relative yield decrease (RYD) probability of exceedance curves, built for each climate region and soils of medium TAW (136-157 mm m -1) are compared in Fig.7a. RYD is the largest in Sandanski ranging from 65 to 85% over the average demand years (40<PRYD<75%). It is also very high in Plovdiv (60<RYD<70%) but lower in Sofia and Pleven (30<RYD<50%) over the same years. In the very dry years (PRYD<5%) yield losses are over 95% in Plovdiv and Sandanski and more than 85% in Sofia field, Silistra, Pleven and Varna (Fig.7a).
Considering a trend of NIR for 1951-2004, an average increase by 80mm is found for Plovdiv; Contrarily, a 19% grain production decrease for rainfed maize is found there.
Maps of High Peak Season SPI2 “July-Aug” spatial distribution relative to the very dry (2000), the average (1970) and the moderately dry (1981) year are elaborated, as presented in Figs. 10a, 10b and 10c. The latter maps and derived relationships (Table 2) are used then to predict the distribution of yield losses of rainfed maize (Figs. 10d, 10e, 10f) or maize irrigation requirements in Bulgaria (Figs.10g, 10h and 10i).
The WinISAREG model (Pereira et al., 2003) is an irrigation scheduling simulation tool for computing the soil water balance and evaluating the respective impacts on crop yields. The model adopts the water balance approach of Doorenbos & Pruitt (1977) and the updated methodology to compute crop evapotranspiration and irrigation requirements (Allen et al. 1998).
٠Considering an economical relative yield decrease (RYD) threshold of 60 and 48% of the potential maize productivity in Plovdiv and Sofia, 30 % of years are risky when TAW=180 mm m-1 in Plovdiv, that is double than in Sofia and half than in Sandanski (Fig.7b). In North Bulgaria the economical RYD threshold is 67, 55 and 60% for Pleven, Lom and Silistra. When TAW=180 mm m-1 only about 10% of the years are risky in Pleven and Silistra that is half than in Lom. When TAW is medium (157 mm m-1) the risky years rise to 19, 35 and 45 % in the three sites respectively and reach 50% in Varna (Fig.7a).
V.Koinov, I. Kabakchiev, K. Boneva, 1998
Fig.10 Spatial distribution of seasonal SPI2 “July-Aug” a)-c)/ Relative yield decrease for rainfed maize (RYD, %) d)-f) / Net irrigation requirements (NIR, mm) g)-i) relative to the year of: a), d) and g) extreme (2000); b), e) and h) average
(1970) and c), f) and i) moderate (1981) irrigation demand, Bulgaria
Map of Bulgaria with experimental
fields of ISSAPPNP
and meteorological
stations of NIMH.
Monthly precipitation in June, July and August relative to the average demand year are the largest in Sofia field, which is double than in Plovdiv, Sandanski, Lom and Varna (Fig.2).
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
PRYD (%)
RY
D (%
)
18 % risky years (9/51) for Pleven 35 % risky years (18/51) for Silistra 39 % risky years(20/51) for Sofia
59 % risky years (30/51) for Plovdiv 82 % risky years(42/51) for Sandanski
49 % risky years(25/51) for Varna
Results and findings:
Fig.2 Average monthly precipitation for the average (probability of exceedance of precipitation P=50%), wet (P=10%) and dry (P=90%) season at: a) Sofia and b) Plovdiv, May-Sept 1951-2004.
RYD
Climate Region
Soil groups according to total available water TAW:
Small116 mm m-1
Medium136-157 mm m-1
Large173-180 mm m-1
Transitional Mediterranean
Sandanski +1.40 +1.00 +0.20
Transitional Continental
Stara Zagora +0.50 +0.10 -0.50
Plovdiv +0.15 0.00 -0.50
Moderate Continental
Lom +0.15 -0.10 -0.75
Sofia 0.00 -0.25 -0.90
Silistra -0.15 -0.50 -1.25
Pleven -0.50 -0.75 -1.50
Northern Black Sea Varna +0.21 -0.21 -1.05
Climate Region
Soil groups according to total available water TAW:
Small116 mm m-1
Medium136-157 mm m-1
Large173-180 mm m-1
Transitional Mediterranean
Sandanski +1.40 +1.00 +0.20
Transitional Continental
Stara Zagora +0.50 +0.10 -0.50
Plovdiv +0.15 0.00 -0.50
Moderate Continental
Lom +0.15 -0.10 -0.75
Sofia 0.00 -0.25 -0.90
Silistra -0.15 -0.50 -1.25
Pleven -0.50 -0.75 -1.50
Northern Black Sea Varna +0.21 -0.21 -1.05