1

The dependence structure between rainfall and

temperature A copula analysis
Rong-Gang Cong
1*

1. Centre for Environmental and Climate Research (CEC), Lund University, Lund, SE-
22362, Sweden.


*Corresponding author:
Dr. Rong-Gang Cong
Email: cascong@126.com; Ronggang.Cong@cec.lu.se











2

Abstract
Quantitative analysis and simulation of rainfall and temperature are extremely
important for agricultural production. However, there may be positive or negative
correlation between rainfall and temperature, especially stronger for some seasons.
Instead of Monte Carlo simulation, five families of copula models are employed to model
the dependence structure between them. The historical weather data of Scania in Sweden
is used as an example. Heteroscedasticity and autocorrelation of sample data are also
considered in the modeling process. Based on the criterion of AIC (BIC), student copula
is found to be most suitable to model the bivariate distribution of rainfall and temperature.
The results can be a good reference for agricultural planning.

Keywords
Agricultural economics; Weather simulation; Multivariate analysis; Copula model; Rainfall;
Temperature

1. Introduction
Weather is the key source of uncertainty affecting crop yield especially in the context
of climate change [1-3]. Vergara et al. studied the potential impact of catastrophic
weather on crop insurance industry and found about 93% crop loss was directly related to
unfavorable weather [4], as shown in Fig. 1. Accurate forecasting of weather 3-6 months
3

ahead of time can potentially allow farmers to do decisions to reduce the risk exposure or
take advantage of expected favorable climate [5].

Fig. 1 Sources of uncertainty affecting crop yield
Data source: AIR Worldwide Corporation
Among variables relevant to weather, rainfall and temperature are two important
factors which have a large effect on crop yield [6-10]. In theory, temperature may affect
the duration of the growing period and rainfall (through soil water availability) may
affect the duration of growth and the production of plant (leaf area and the photosynthetic
efficiency) [11, 12]. There is quite a lot of literature studying the effect of temperature
and rainfall on crop yield. In respect of the impact of rainfall on crop yield, using lentil in
North Africa and West Asia as an example, Erskine and Ashkar quantified the effect of
rainfall on lentil seed yield and found that rainfall accounted for 79.8% of the variance of
seed yield [13]. Lobell et al. further studied the relationship between crop yield and
rainfall for 12 major Californian crops and found rainfall was able to explain more than
7%
37%
33%
13%
5%
1%
4%
93%
Other perils Drought & Heat
Excess Moisture Hail
Cold, Frost & Freeze Flood
Directly
related to
weather
4

60% of the observed variability in yields for most crops [14]. Cooper et al. investigated
the rain-fed farming system of sub-Saharan Africa and found that not only the seasonal
rainfall totals and their season-to-season variability were in themselves important, but
also the within season variability had a major effect on crop productivity [15].
In respect of the impact of temperature on crop yield, Muchow et al. examined the
effect of temperature on potential maize yield and found that temperature primarily
affected growth duration with lower temperature increasing the length of time that the
crop could intercept radiation [16]. Cane et al. also examined the impact of sea surface
temperature on maize yield and found that a strong correlation between temperature and
maize yields in Zimbabwe [17]. Lobell and Asner studied the climate contribution to U.S.
agricultural yield and found a roughly 17% relative decrease in both corn and soybean
yield for each degree increase in growing season temperature [18]. In summary, rainfall
and temperature have been considered to be two most important climatic factors affecting
agricultural production [19-21].
Since temperature and rainfall have a prominent impact on crop yield, accurate
simulation of temperature and rainfall is important not only for meteorology but also for
agricultural economics. However, in reality it is difficult to use traditional probability
methods to simulate rainfall and temperature independently due to the dependence
structure between them [22-24]. From the perspective of space, it is generally believed
that there are positive (negative) correlation between rainfall and temperature over
tropical oceans (lands) [25]. Aldrian and Susanto examined the relationship between
rainfall and sea surface temperature, and found that Indonesian rainfall variability
revealed some sensitivity to sea-surface temperature variability in adjacent parts of the
5

Indian and Pacific Oceans [26]. Black also studied the relationship between Indian Ocean
sea surface temperature and East Africa rainfall, and concluded that strong East African
rainfall was associated with warming in the Pacific and Western Indian Oceans and
cooling in the Eastern Indian Ocean [27]. From the perspective of time, it is generally
believed that the correlation between rainfall and temperature differs in different months.
Rajeevan et al. examined the temporal relationship between land surface temperature and
rainfall [28]. They found that temperature and rainfall were positive correlated during
J anuary and May but negatively correlated during J uly. From the perspective of annual
data, Huang et al. also found a negative correlation between rainfall and temperature in
Yellow River basin of China [29].
If not considering the dependence structure between rainfall and temperature, the most
direct method to simulate random climatic variables may be the Monte Carlo simulation
[30]. However, Monte Carlo simulation is not good at modeling the serial correlation. In
this respect, Copulas provide an alternative way to model joint distributions of random
variables with greater flexibility both in terms of marginal distributions and the
dependence structure [31]. Copulas have been used in financial literature for quite some
time [32-34], but have not made their way yet to the literature about agricultural weather
simulation quite often. In the respects of temperature and rainfall, AghaKouchak et al.
applied two different elliptical copula families, namely, Gaussian and t-copula, to
simulate the spatial dependence of rainfall and found that using t-copula might have
significant advantages over the well-known Gaussian copula particularly with respect to
extremes [35]. Serinaldi also studied the spatial dependence of rainfall and found positive
contemporaneous pairs of rainfall observations, which described the inter-site
6

dependence [36]. Laux et al. highlighted the importance of pretreatment of
meteorological data in the Copula modeling process [37]. Laux et al. used the Clayton
copula to construct the bivariate distribution of drought duration and drought intensity
[38]. Similar applications of Clayton copula can also be found in the studies of Favre et al.
and Shiau et al. [39, 40]. However, they also raised the question as to which copula
model fitted best to the empirical data. The only literature about application of Copula
simulation to model the inter-dependence between temperature and rainfall up to now is
Scholzel et al. [41]. They used a simple statistical model based on the copula approach to
describe the phenomenon that cold periods were accompanied by small precipitation
amounts.
Inspired by Dupuiss study on hydrological random variables [42], the purpose of the
present paper is to illustrate the pretreatment process of meteorological data, demonstrate
the application of different copulas to modeling of joint distributions of rainfall and
temperature, and select the most suitable copula function according to some criterions.
The rest of the paper is organized as follows. The next section presents materials and
methods used in this paper. Section 3 illustrates the results from estimating the joint
distribution of temperature and rainfall, and evaluates five copulas performances
according to some criterions. The last section presents concluding remarks.
7

2. Materials and methods
2.1 Study area

Fig. 2. The geographical position and real scenery of Scania, Sweden
Data resource: http://www.visitsweden.com and http://en.wikipedia.org/wiki/Scania
Scania is Swedens southernmost region (Fig. 2) and one of Northern Europes richest
farming districts. As Scania is surrounded by water on three sides (the Baltic Sea, the
Kattegat Sea and the resund sound), it has a maritime climate, especially along the
south and east coasts. The winters are mild (few days of snow), but the summers are not
warmer than them in the rest of southern Sweden.
8

2.2 Data collection and preliminary analysis
Monthly temperature and rainfall data in Scania from 1961 to 2010 is collected from
Swedish Meteorological and Hydrological Institute.
(1) Temperature

Fig. 3. Monthly temperature in Scania, Sweden from 1961 to 2010
Monthly temperature in Scania has a similar trend from 1961 to 2010, as shown in
Fig.3. The temperature reaches its peak in J uly and its bottom in February respectively.
From April to November, the temperature is always greater than 0. It is also found that
the variability of temperature in J anuary and February is relatively large. Some
descriptive statistics are listed in Table 1.
Table 1 Descriptive statistics for monthly temperature from 1961 to 2010 (unit: )
J anuary February March April May J une
Maximum 5.10 5.20 5.80 8.30 13.10 18.00
Minimum -5.20 -5.20 -2.30 1.90 8.40 12.00
Mean 0.83 0.54 2.21 5.68 10.58 14.78
Standard
deviation
2.48 2.39 1.85 1.36 1.20 1.07
9

Variation
coefficient
2.99 4.44 0.84 0.24 0.11 0.07
J uly August September October November December
Maximum 21.00 21.50 16.90 13.20 8.30 7.10
Minimum 13.90 14.60 11.60 7.60 2.70 -2.80
Mean 16.99 17.04 13.98 10.03 5.80 2.49
Standard
deviation
1.59 1.45 1.20 1.33 1.27 1.88
Variation
coefficient
0.09 0.09 0.09 0.13 0.22 0.76

(2) Rainfall
Compared with temperature, monthly rainfall in Scania doesnt show a clear trend from
1961 to 2010. From J une to November, the average rainfalls are relatively high (Fig. 4).
Some descriptive statistics are listed in Table 2.

Fig. 4. Monthly rainfall in Scania, Sweden from 1961 to 2010
Table 2 Descriptive statistics for monthly rainfall from 1961 to 2010 (Unit: mm)
J anuary February March April May J une
10

Maximum 70.00 50.00 73.50 87.90 90.60 123.3
Minimum 1.00 5.00 3.30 3.80 6.30 0.1
Mean 35.19 25.14 30.07 32.20 39.74 46.28
Standard
deviation
17.07 11.38 16.25 18.86 20.46 26.82
Variation
coefficient
0.49 0.45 0.54 0.59 0.51 0.58
J uly August September October November December
Maximum 147.60 189.90 161.90 106.30 95.00 106.00
Minimum 7.40 5.70 7.30 4.50 17.00 4.80
Mean 51.15 58.33 49.20 45.90 45.73 40.80
Standard
deviation
30.76 39.97 31.48 24.26 19.65 19.03
Variation
coefficient
0.60 0.69 0.64 0.53 0.43 0.47

(3) The relationship between rainfall and temperature
The physical rationale behind the relationship between rainfall and temperature is that
rainfall may affect soil moisture which may in turn affect surface temperature by
controlling the partitioning between the sensible and latent heat fluxes [43]. Here the
Pearson method is employed to calculate the correlation coefficient between monthly
rainfall and temperature. It is found that there is a significant negative correlation
between rainfall and temperature from April to J uly and September (at the 5% confidence
level), and a positive correlation in J anuary (at the 10% confidence level) (Table 3).
Table 3 Correlation analysis for monthly temperature and rainfall from 1961 to 2010
J anuary February March April May J une
Pearson
correlation
coefficients
0.27 0.19 0.08 -0.46 -0.43 -0.30
P value 0.06 0.20 0.59 0.001 0.002 0.04
J uly August September October November December
Pearson
correlation
coefficients
-0.43 -0.03 -0.29 -0.15 0.01 0.14
P value 0.002 0.82 0.04 0.30 0.95 0.33
11


2.3 Methods
Here we use the Copula functions to model the dependence between the probability
distributions of a certain month temperature and rainfall. Let X and Y be continuous
random variables representing temperature and rainfall, with cumulative distribution
functions ( ) Pr( )
X
F x X x = and ( ) Pr( )
Y
G y Y y = , respectively. Following Sklar [44],
there is a unique function C such that:
Pr( , ) ( ( ), ( )) X x Y y C F x G y = (1)
where ( , ) Pr( , ) C u v U u V v = is the distribution of the pair ( , ) ( ( ), ( )) U V F X G Y = whose
margins are uniform on [0, 1]. The function C is called a copula. As argued by J oe [45] or Nelsen
[46] among others, C characterizes the dependence in the pair (X, Y). There are many parametric
copula families available, which usually have parameters that control the strength of dependence.
Among them, five families of copulas commonly used are considered. They are listed in table 4,
along with their parameter range. The first three are Archimedean [45] and the last two are meta-
elliptical [47].
Table 4. Five families of copulas
Family C(u,v) Range of
Normal
1 1
( ( ), ( )) N u v

[-1,1]
Student
1 1
,
( ( ), ( )) T T u T v



[-1,1]
Clayton
1/
( +v 1) u

(0, )
Frank
1
ln{1 ( 1)( 1) / ( 1)}
u v
e e e

+
( ) { } , 0
12

Gumbel
{ }
1/
exp (ln ln ) u v


+
[1, ]
Source: [48]
After calculating the parameter for each copula, it is necessary to decide which family is the
best representation of the dependence structure between variables. There are a few techniques to
select the best copula. One of them is based on distance measures pertaining to candidate models
(copulas) distributions and the empirical datas distribution [48, 49]. Other alternative methods
include Likelihood Ratio tests and approaches related to information criteria [33], such as Akaike
and Schwarz's Bayesian Information Criteria which are adopted because they can describe the
tradeoff between bias (or accuracy) and variance (complexity) in model construction. The Akaike
information criterion (AIC) is a measure of the relative goodness of fit of a statistical model. Its
definition is as follows:
2 2ln( ) AIC k L = (2)
In equation (2), k is the number of parameters in the copulas, and L is the maximized value of
the likelihood function for the copulas. The Bayesian information criterion (BIC) was developed
by Schwarz using Bayesian formalism. Its definition is as follows:
=-2ln( )+ln( ) BIC L k N (3)
In equation (3), N is the sample size.
3. RESULTS AND DISCUSSION
Temperature and rainfall data in April from 1961 to 2010 is employed as an example (Fig. 5).
There is significant negative relationship (Pearson correlation coefficient is -0.46, P-value=0.001)
between temperature and rainfall in April. The temperature has negative skewness (-0.35) and the
rainfall has positive skewness (1.07), which may create a heteroscedasticity problem in fitting the
13

model [50]. Inspired by Kim and Ahn [51], the temperature and rainfall data are log-transformed
to remove this effect at first. The logarithmic transformation for the data is invertible, which will
not affect the fitting results.

Fig. 5. Temperature and rainfall in April from 1961 to 2010
Following Benth and Saltyte-Benths instructions [52], the time series of temperature and
rainfall are tested for autocorrelation using the Q-statistics (Fig. 6). Autocorrelation describes the
correlation between values of temperature (or rainfall) at different points in time, as a function of
the time difference. The presence of autocorrelation increases the variances of residuals and
estimated coefficients, which reduce the models efficiency. The LjungBox Q test is a type of
statistical test of whether autocorrelations of a time series are different from zero [53]. The Q-
statistics is defined as follows:
2
1

=( +2)
h
k
k
p
Q n n
n k
=

(4)
where n is the sample size,
2

k
p is the sample autocorrelation at lag k, and h is the number of lags
being tested. It is found the first order autocorrelations are strong both for temperature (Q-
stat=6.32, P-value=0.01) and rainfall (Q-stat=4.52, P-value=0.03), as shown in Fig. 6.
1 2 3 4 5 6 7 8 9
0
20
40
60
80
100
Temperature
R
a
i
n
f
a
l
l
14


Fig. 6. Sample autocorrelation function (ACF) of the temperature and rainfall in April
Therefore, AR(1) model is used to eliminate the autocorrelation in the series as shown in
equation (5).
1
** **
1
** **
0.48 0.35
(4.7 ) (2.56 )
1.85 0.29
(9.06 ) ( 2.1 )
t t t
t t t
tempe tempe
rain rain

= + +
= +

(5)

Note: The numbers in the bracket are t-values.
**
stands for the statistical significance in the 95%
confidence level.
Residuals
t
and
t
are tested where only weak autocorrelations are found (Fig. 7).
15


Fig. 7 Sample autocorrelation function (ACF) of AR adjusted temperature and rainfall in April
Except for autocorrelation, time trends are also found in the series of
t
and
t
. Based Manton
et al.s research [54], the time trends should be removed from the series to obtain a stationary
process. The detrend functions are as shown in equation (5).
** **
** **
0.08 0.0032
( 2.65 ) (3.04 )
0.17 0.007
(2.3 ) ( 2.65 )
t t
t t
t
t


= + +

= +

(5)

It is found that temperature has an increasing trend and rainfall has a decreasing trend from
1961 to 2010 (Fig. 8). The annual rate of increase (decrease) in temperature (rainfall) is 0.0032
(0.007 mm).
t
rtempe and
t
rrain are used to represent
t
and
t
respectively and shown in Fig. 9.
0 2 4 6 8 10 12 14 16 18 20
-0.5
0
0.5
1
Lag
S
a
m
p
l
e

A
u
t
o
c
o
r
r
e
l
a
t
i
o
n
Sample ACF of temperature residuals in April
0 2 4 6 8 10 12 14 16 18 20
-0.5
0
0.5
1
Lag
S
a
m
p
l
e

A
u
t
o
c
o
r
r
e
l
a
t
i
o
n
Sample ACF of rainfall residuals in April
16


Fig. 8. Residuals for AR adjusted temperature and rainfall in April

Fig. 9. Scatters of residuals for detrend adjusted temperature and rainfall in April
Based on the inference for the margins (IFM) [55], estimations and related evaluation index of
Copulas for
t
rtempe and
t
rrain are calculated as shown in Table 5.
Table 5 Results of different copula model for temperature and rainfall in April
Normal Student Clayton Frank Gumbel
-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
rtempe
r
r
a
i
n
17

-0.34 -0.31 0.001 0.001 1.1
log-likelihood 3.05 4.11 -0.0007 -0.0002 -1.86
AIC -6.06 -8.15 0.042 0.041 3.75
BIC -6.02 -8.07 0.081 0.08 3.79

The log-likelihood of Student Copula is largest while the AIC and BIC for Student Copula are
smallest, which means that the student copula model may be the most suitable. The comparison
of real and simulated temperature and rainfall is shown in Fig. 10.

Fig. 10 Real residuals (above) and Student-based Copula simulated (below) residuals scatters
-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
rtempe
r
r
a
i
n
-0.3 -0.2 -0.1 0 0.1 0.2 0.3
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
x
y
18

The purpose of this paper is to simulate the pair of rainfall and temperature using the copula
model. Therefore, the accuracy of model is utmost important. Although Table 5 has given some
statistical tests and Fig. 10 has given some straight evidences, it is necessary to use the contours
of cumulative distribution functions to show the dependence structure of real data and simulated
data.
The contours of cumulative distribution functions (CDF) for empirical data and five copula
models are plotted to visualize the difference as shown in Fig. 11-15. It is also found that Student
copula model fits the empirical data best.

Fig. 11. Empirical vs Gauss Fitted CDF

Fig. 12. Empirical vs Student Fitted CDF
0
.
2
0
.
2
0.2
0
.
4
0.4
0
.
6
0.6
0
.
8


-0.25 -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0
.
2
0
.
2
0.2
0
.
4
0
.4
0
.
6
0.6
0
.
8
rtempe
r
r
a
i
n
-0.25 -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
19


Fig. 13. Empirical vs Clayton Fitted CDF

Fig. 14. Empirical vs Frank Fitted CDF

Fig. 15. Empirical vs Gumbel Fitted CDF
Note: The dot lines are empirical CDFs while the solid lines are the simulated CDFs
0
.
2
0
.
2
0.2
0
.
4
0
.4
0
.
6
0.6
0
.
8
rtempe
r
r
a
i
n
-0.25 -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0
.
2
0
.
2
0.2
0
.
4
0.4
0
.
6
0.6
0
.
8
rtempe
r
r
a
i
n
-0.25 -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0
.
2
0
.
2
0.2
0
.
4
0.4
0
.
6
0.6
0
.
8
rtempe
r
r
a
i
n
-0.25 -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
20

4. CONCLUSIONS
This paper presents a copula-based methodology for modeling the joint distribution of
temperature and rainfall, which are utmost important for agricultural production especially in the
context of climate change. Copulas have been used extensively in financial literature, but have
not been widely used in agricultural economics. The copula approach provides a powerful and
flexible method to model multivariate distributions and thus go beyond joint normality,
regressibility, and mean-variance criterion. Accurate estimation of weather may help to improve
the risk management in agricultural planning.
A shortcoming of copula method is the arbitrariness of the copula selection. This study taken in
Scania is held to construct a general copula framework to model to dependence structure of
rainfall and temperature. Different from previous study by Schoelzel et al. [41], it compares
different copula forms and shows how to select the optimal copula based on some statistical tests.
Its advantage is that it does not require any assumptions and is primarily data driven thus
minimizing the subjectivity introduced by the researcher. The empirical results show that
different from other application of copula models, student copula model has its advantage to
model the dependence structure between rainfall and temperature.
Further research in this area needs to be connected with the analysis of crop production in
agricultural economics, such as the relationship among temperature, rainfall and crop yield. It will
be the basis for future study on agricultural economics.
Acknowledgements
This research is funded by Biodiversity and Ecosystem services in a Changing
Climate BECC and SAPES. The authors gratefully acknowledge the good research
environments provided by the Centre for Environmental and Climate Research (CEC),
21

Lund University and AgriFood Economics Centre. We thank Katarina Hedlund and
Henrik Smith for fruitful discussions. We also thank the editor and anonymous referees
for their helpful suggestions and corrections on the earlier draft of our paper according to
which we improved the content.
Reference:
1. Black J R, Thompson SR (1978) Some Evidence on Weather-Crop-Yield Interaction. Am J
Agri Econ 60: 540-543.
2. Alexandrov VA, Hoogenboom G (2000) The impact of climate variability and change on crop
yield in Bulgaria. Agri For Meteorol 104: 315-327.
3. Chloupek O, Hrstkova P, Schweigert P (2004) Yield and its stability, crop diversity,
adaptability and response to climate change, weather and fertilisation over 75 years in the
Czech Republic in comparison to some European countries. Field Crop Res 85: 167-190.
4. Vergara O, Zuba G, Doggett T, Seaquist J (2008) Modeling the Potential Impact of
Catastrophic Weather on Crop Insurance Industry Portfolio Losses. Am J Agri Econ 90:
1256-1262.
5. J ones J W, Hansen J W, Royce FS, Messina CD (2000) Potential benefits of climate forecasting
to agriculture. Agri Ecosyst Environ 82: 169-184.
6. Runge ECA (1968) Effects of Rainfall and Temperature Interactions During the Growing
Season on Corn Yield. Agron J 60: 503-507.
7. Abbate PE, Dardanelli J L, Cantarero MG, Maturano M, Melchiori RJ M, Suero EE (2004)
Climatic and water availability effects on water-use efficiency in wheat. Crop Science
44(2): 474-483.
8. Calderini DF, Abeledo LG, Savin R, Slafer GA (1999) Effect of temperature and carpel size
during pre-anthesis on potential grain weight in wheat. J ournal of Agricultural Science
132(4): 453-459.
9. Medori M, Michelini L, Nogues I, Loreto F, Calfapietra C (2012) The impact of root
temperature on photosynthesis and isoprene emission in three different plant species. The
Scientiric World J ournal 2012: 525827.
22

10. Passarella VS, Savin R, Slafer GA. (2008) Are temperature effects on weight and quality of
barley grains modified by resource availability? Australian J ournal of Agricultural
Research 59(6): 510-516.
11. Cantelaube P, Terres J M (2005) Seasonal weather forecasts for crop yield modelling in
Europe. Tellus A 57(3): 476-487.
12. Olesen J E, Bindi M (2002) Consequences of climate change for European agricultural
productivity, land use and policy. European J ournal of Agronomy 16: 239-262.
13. Erskine W, El Ashkar F (1993) Rainfall and temperature effects on lentil ( Lens culinaris)
seed yield in Mediterranean environments. J Agri Sci 121: 347-354.
14. Lobell DB, Cahill KN, Field CB (2007) Historical effects of temperature and precipitation on
California crop yields. Climatic Change 81: 187-203.
15. Cooper PJ M, Dimes J , Rao KPC, Shapiro B, Shiferaw B, Twomlow S (2008) Coping better
with current climatic variability in the rain-fed farming systems of sub-Saharan Africa:
An essential first step in adapting to future climate change? Agriculture, Ecosystems and
Environment 126: 2435.
16. Muchow RC, Sinclair TR, Bennett J M Temperature and Solar Radiation Effects on Potential
Maize Yield across Locations. Agron J 82: 338-343.
17. Cane MA, Eshel G, Buckland RW (1994) Forecasting Zimbabwean maize yield using eastern
equatorial Pacific sea surface temperature. Nature 370: 204-205.
18. Lobell DB, Asner GP (2003) Climate and Management Contributions to Recent Trends in
U.S. Agricultural Yields. Science 299(5609): 1032.
19. Lobell DB, Field CB (2007) Global scale climatecrop yield relationships and the impacts of
recent warming. Environmental Research Letters 2(1): 014002.
20. Kaufmann RK, Snell SE (1997) A Biophysical Model of Corn Yield: Integrating Climatic and
Social Determinants. Am. J . Agr. Econ. 79(1): 178-190.
21. Riha SJ , Wilks DS, Simoens P (1996) Impact of temperature and precipitation variability on
crop model predictions. Climatic change 32(3): 293-311.
22. Shukla J , Misra BM (1977) Relationships between sea surface temperature and wind speed
over the central Arabian Sea, and monsoon rainfall over India. Mon weather rev 105:
998-1002.
23. Moise AF, Colman RA, Brown J R (2012). Behind uncertainties in projections of Australian
tropical climate: Analysis of 19 CMIP3 models. J ournal of geophysical research 117:
D10103.
23

24. Tanarhte M, Hadjinicolaou P, Lelieveld J (2012) Intercomparison of temperature and
precipitation data sets based on observations in the Mediterranean and the Middle East.
J ournal of Geophysical Research D: Atmospheres 117(12): D12102.
25. Adler RF, Gu G, Wang J J , Huffman GJ , Curtis S, Bolvin D (2008). Relationships between
global precipitation and surface temperature on interannual and longer timescales (1979
2006). J ournal of geophysical research 113: D22104.
26. Aldrian E, Dwi Susanto R (2003) Identification of three dominant rainfall regions within
Indonesia and their relationship to sea surface temperature. Int J Climato 23: 1435-1452.
27. Black E (2005) The relationship between Indian Ocean seasurface temperature and East
African rainfall. Phlio Tran R Soc A 363: 43-47.
28. Rajeevan M, Pai DS, Thapliyal V (1998) Spatial and Temporal Relationships Between Global
Land Surface Air Temperature Anomalies and Indian Summer Monsoon Rainfall.
Meteorol Atmos Phys 66: 157-171.
29. Huang Y, Cai J , Yin H, Cai M (2009) Correlation of precipitation to temperature variation in
the Huanghe River (Yellow River) basin during 19572006. J ournal of Hydrology 372(1-
4): 1-8.
30. Rahman A, Weinmann PE, Hoang TMT, Laurenson EM (2002) MonteCarlo simulation of
flood frequency curves from rainfall. J ournal of Hydrology 256(3-4): 196-210.
31. AghaKouchak A, Bardossy A, Habib E (2010) Conditional simulation of remotely sensed
rainfall data using a non-Gaussian v-transformed copula. Advances in Water Resources
33(6): 624-634.
32. Malevergne Y, Sornette D (2003) Testing the Gaussian copula hypothesis for financial assets
dependences. Quan Finan 3: 231-250.
33. Patton AJ (2009) CopulaBased Models for Financial Time Series. Handbook of Financial
Time Series. Springer Berlin Heidelberg: 767-785.
34. Genest C, Gendron M, Bourdeau-Brien M (2009) The Advent of Copulas in Finance. Euro J
Finan 15: 609-618.
35. AghaKouchak A, Bardossy A, Habib E (2010) Copula-based uncertainty modelling:
application to multisensor precipitation estimates. Hydrol Process 24: 2111-2124.
36. Serinaldi F (2008) Analysis of inter-gauge dependence by Kendalls sK, upper tail
dependence coefficient, and 2-copulas with application to rainfall fields. Stoch Environ
Res Risk Assess 22: 671-688.
24

37. Laux P, Vogl S, Qiu W, Knoche HR, Kunstmann H (2011) Copula-based statistical
refinement of precipitation in RCM simulations over complex terrain. Hydrol Earth Syst
Sci 15: 24012419.
38. Laux P, Wagner S, Wagner A, J ocobeit J , Bardossy A, Kunstmann H (2009) Modelling daily
precipitation features in the Volta Basin of West Africa. Int J Climatol 29: 937-954.
39. Favre AC, El Adlouni IS, Perreault L, Thiemonge N, Bobee B (2004) Multivariate
hydrological frequency analysis using copulas. Water Resources Research 40: 1-12.
40. Shiau J -T, Feng S, Nadarajah S (2007) Assessment of hydrological droughts for the Yellow
River, China, using copulas. Hydrological Processes 21(16): 21572163.
41. Schoelzel C, Friederichs P (2008) Multivariate non-normally distributed random variables in
climate research introduction to the copula approach. Nonlin Processes Geophys 15:
761-772.
42. Dupuis DJ (2007) Using Copulas in Hydrology: Benefits, Cautions, and Issues. J Hydrol Eng
12: 381-393.
43. Huang J , Van Den Dool HM (1993) Monthly precipitation-temperature relations and
temperature prediction over the United States. J Clim 6(6): 1111-1132.
44. Sklar A (1973) Random variables, joint distribution functions, and copulas. Kybernetika 9:
449-460.
45. J oe H (1997) Multivariate Models and Dependence Concepts. London: Chapman and Hall.
46. Nelsen R (1999) An Introduction to Copulas. New York: Springer.
47. Fang HB, Fang KT, Kotz S (2002) The Meta-elliptical Distributions with Given Marginals. J
Mult Anal 82: 1-16.
48. Gregoire V, Genest C, Gendron M (2008) Using copulas to model price dependence in energy
markets. Energy Risk 5: 58-64.
49. Kole E, Koedijk K, Verbeek M (2007) Selecting copulas for risk management. J Banking
Finan 31: 2405-2423.
50. Katircioglu ST. RESEARCH METHODS IN BANKING AND FINANCE.
http://www.emu.edu.tr/salihk/courses/bnfn504/chp-11.pdf
51. Kim T-W, Ahn H (2009) Spatial rainfall model using a pattern classifier for estimating
missing daily rainfall data. Sto Environ Res Risk Assess 23: 367-376.
52. Benth FE, altyt-Benth J (2005) Stochastic Modelling of Temperature Variations with a
View Towards Weather Derivatives. Applied Mathematical Finance 12: 53-85.
53. Ljung GM, Box GEP (1978) On a Measure of Lack of Fit in Time Series Models. Biometrika
65(2): 297-303.
25

54. Manton MJ , Della-Marta PM, Haylock MR, Hennessy KJ , Nicholls N, et al. (2001) Trends in
extreme daily rainfall and temperature in Southeast Asia and the South Pacific: 1961
1998. Int J Climato 21: 269-284.
55. J oe H, Xu J J (1996) The Estimation Method of Inference Function for Margins for
Multivariate Models. Department of Statistics, University of British Columbia.