Analyzing the Impact of Climate Change on Cotton Yield Using Spatial Analysis and Statistical Modeling in the Indus River Basin, Pakistan

Naveed M; He HS; Yasir QM; Du H; Satti Z

Research Article

Ann Agric Crop Sci. 2021; 6(5): 1089.

Analyzing the Impact of Climate Change on Cotton Yield Using Spatial Analysis and Statistical Modeling in the Indus River Basin, Pakistan

Naveed M¹, He HS²*, Yasir QM¹, Du H¹ and Satti Z¹

¹Key Laboratory of Geographical Processes and Ecological Security of Changbai Mountains, Ministry of Education, School of Geographical Sciences, Northeast Normal University, Changchun, People’s Republic of China

²School of Natural Resources, University of Missouri, Columbia, MO, USA

*Corresponding author: Hong S He, School of Natural Resources, University of Missouri Columbia, MO 65203, USA

Received: June 28, 2021; Accepted: July 23, 2021; Published: July 30, 2021

Abstract

Cotton is an excessive growing fiber crop in the world. In Pakistan, climate warming is hurting the cotton crop. To identify countermeasures of climate change impacts on crop production, it is needed to explore the changes in crop yield and their relationship to climate change. This study, we conducted spatial analysis of cotton yield and climate data in the Indus River basin in Pakistan from 1989- 2018. We analyzed the trend in cotton production indices and climate variables using the Mann-Kendal nonparametric test, and the impact of climate change on cotton yield through Pearson’s correlation and best-subset multivariate linear regression model. Sunshine hours positively impacted on cotton yield, but an increasing trend in average temperature, minimum temperature and air relative humidity in main crop areas had a negative impact on the cotton crop. Our results illustrate the relationship between the cotton yield and climate change variables on regional basis in the Indus River basin, Pakistan. This research could provide a reference for understanding the intolerance of cotton in the Indus River basin, Pakistan to climate change.

Keywords: Climate change; Cotton; GIS; Multivariate regression

Abbreviations

AIC: All-Subset Information Criteria; BIC: Bayesian Information Criteria; GIS: Geographic Information System; MMK: Modified Mann-Kendal; MLR: Multivariate Linear Regression; RMSE: Root Mean Square Error

Introduction

Cotton is one of the major cash crops worldwide and the natural fiber crop in textile industries [1,2]. Cotton is also fine chemical raw material and an oil crop [3] planted in more than 80 countries around the world [4]. Cotton cultivation is thoroughly linked to climatic conditions and irrigation and greatly influences total nitrogen emissions during the growing period [5]. Thus, timely and efficient data of the spatial distribution of cotton is vital for agricultural management, regional crop response, and agricultural policymaking under climate change [6]. Cotton is successfully grown at temperatures ranging from 28.20°C in China to 41.80°C in Sudan. Seasonally, average temperatures are about 36.80°C in the Indus River basin, Pakistan. However, history exhibits that heat stress is a major obstacle to cotton growth in some states, including Pakistan, Syria and India. Moreover, the increase in temperature can hurt the budget of cotton in regions where it is already grown at a temperature of almost 40°C [7]. Pakistan is fourth largest cotton producer in the world [8]. Indus River basin cotton belt covers the 88% of the cotton crop area and 90% of the cotton production of Pakistan (Pakistan Bureau 2018), which has low seasonal rainfall and high temperatures.

Climate change has been one of one of the most important issues affecting many systems [9]. It is estimated that anthropogenic activities caused about 1°C global warming at pre-industrial levels [10]. Under various emission scenarios, the projected increase ranges from 2°C to 5°C by the end of the twenty-21 century [11]. The global climate system and agroecology have already and will continue to analyzing impact of warming, e.g., [12,13]. For the past several decades, scientists have been investigating the influences of climatic variations on the cotton yield. Daily minimum and maximum temperature have been examined that extreme weather events have adverse effects on ecological systems [14]. Rosenzweig et al. studied that climate change is expected to increase the vulnerability of agricultural systems by changes in rainfall patterns. Increasing temperature and increased frequency of extreme weather events in most parts of the world [10] especially in Pakistan [15,16]. Arid to semi-arid regions [15,17-19] are expected to have a negative impact on crops due to projected changes. Increases in temperature [20] and changes in precipitation have a negative impact on cotton growth and productivity [21,22]. The optimum temperature for efficient growth is reported to be 33°C, whereas substantial reduction in flower and boll retention has been recorded when temperature is above 36°C [16,23,24]. Variation in temperature have a significant impact on cotton growth and seed cotton production [25,26]. The boll growth period shortens with high temperatures, resulting in a smaller boll and reduced cotton output [27,28]. Temperature and air humidity are the main factors that affect cotton growth [29]. Using statistical approaches, researchers from different countries investigated the influence of climate change on cotton crop. Tang Xiang-ling and Xin [30] used Multivariate Linear Regression (MLR) analysis to analyze climate and cotton yield variations in Shiheze, China. They have found that cotton production has been improved by an increase in temperature, sunshine, and precipitation. Chao et al. used a provincial-based study and observed that variations in temperature, air and precipitation had a positive effect on cotton productivity. Turner et al. observed that yield of cotton is directly influenced by temperature. Deng Zhenyong et al. [29] analyzed the impact of global warming on various crops in the northwest region of China and found that global warming had a major impact on crop yield as it enhances the production of thermophilic crops (cotton) and changes the growth period of the crops. These adaptive modifications support crops gain additional heat and light for photosynthesis thus helps in the aggregation of dry matter [29]. Moreover, irregular spatiotemporal distributions of precipitation may adversely affect cotton yield [31].

However, the actual climate change varies significantly among regions, which can greatly affect a crop yield. Furthermore, climate conditions are different on different sites in the same region that affect the crop yield [32]. Therefore, the impact of climate change on cotton yield at different locations should be considered to make efficient use of climate resources at specific sites, which can guide to improve plant structures and yield management. Most of the previous studies investigating the impact of climate change on crop production have been focused on grain crops such as rice, corn, and wheat. Although some researchers have examined the influences of climate change on cotton growth and production, the majority of these studies were based on a single site [30,32,33], with very few including multiple locations at regional scales [34]. In addition, these studies did not examine the impact of multiple climate factors on cotton yield based on an assessment of simple correlation and multiple linear regressions.

Pakistan has an arid to semiarid climate with great variability in temperature, with diversified topography and an extremely variable climate. The study of vegetation and moisture stress clearly shown that stress on soil moisture increased in Kalat region of Pakistan that ultimately leads to aridity [9,35]. Pakistan has a diverse and extensive agricultural resource base that spans multiple ecological and climatic zones, with substantial potential for producing a variety of foods and fibers [35]. The cotton cultivation belt along the Indus river irrigation system was recognized as the backbone of the economy for the country [36]. Cotton is commonly known as “white gold,” an important cash crop for Pakistan and grows typically in agricultural plains of the Indus River basin. It contributes 18.5% of the value-added share in national agriculture and about 4.5% to GDP; around two-thirds of the country’s export earnings come from the cotton made-up and textile, which adds over $3.4 billion to the national economy; while hundreds of ginning factories and textile mills in the country depends upon cotton. Primarily millions of farmers are dependent on this crop, in addition to millions of people employed along the entire cotton value chain, from weaving to textile and garment exports. Climate warming has greatly affected agricultural activities in Pakistan. An increase in temperature observed in the country, especially during the last few decades it is very sharp [37]. This higher temperature rises and other alterations in climatic variables have influenced the development of Pakistan’s cotton [7].

In this study, we intend to assess the influence of climatic changes on yield in the Indus river basin, Pakistan. The purpose of this study is to: (1) Analyze changes in climate variables and cotton yield indicators in various sites of Indus River basin, Pakistan, from 1989 to 2018; (2) Describe the relationship and impact of climate variables on cotton yield. This study helps to understand the influence of climatic variability on cotton yield at 12 different sites of the Indus River basin.

Methods

Study area and datasets

Southern Pakistan, specifically the cotton belt area in the Indus River basin, is an important physiographic unit that almost produces 85% of the total cotton crop of Pakistan; and also plays a vital role in the economy of the country (Figure 1). The entire region covers approximately 142075km², which is 18% of the area of Pakistan. Geographically, it extends from 24°56’48” to 31°57’31” North latitude and 69°13’5” to 72°55’41” East longitude, agricultural activities mostly depend upon irrigation [35]. Physically, soil changes from clay loam to sandy with clay dominant towards the south, in terms of elevation, while going from north to south, the elevation decreases. The study area lies in humid, hot, and dry regions where cotton is cultivated [9]. The uppermost temperatures are measured up to maximum from 45°C to 50°C. Because of very short rainfall (155 to 755 mm), cotton production is relying on irrigation as the cotton plants need less water compared to sugarcane and rice.

Figure 1: Study area map indicating the crop and metrological data collection center in Indus River basin, Pakistan.

    
    
    Figure 1: Study area map indicating the crop and metrological data collection center in Indus River basin, Pakistan.

The monthly cotton-growing season (May-September) weather data from 1989 to 2018 at 12 meteorological centers in the Indus River basin were collected from the regional offices of the Pakistan Meteorological Department (http://www.pmd.gov.pk) and the cotton index data of the selected stations were collected from Pakistan Bureau of Statistics (http://www.pbs.gov.pk). The climate variables included T_max, T_min, T_ave, Pre, sunshine duration, and air relative humidity, which is the variable in space and time that may affect the yield of cotton, verifying the need to examine the influence of climate on cotton production [38]. In the Indus River basin, cotton is planted at the end of April and the start of May and cropped once a year. After 6 days of the plantation, seedlings appear; budding begins from the end of June; cracking starts in the middle of July to the start of August; and finally, growth stops at the end of September. May to September is a cotton-growing season that covers 150-170 days in Table 1.

Table 1: The average cotton index, climate variables of cotton growth season (May-Sep) and site information in Indus River basin Pakistan.




  
    Site Name
    Latitude (°)
    Longitude (°)
    Altitude (m)
    Climate variable
    Cotton Index
  
  
    Tmax
    Tmin
    Tave
    Pre
    Sun
    Hum
    Area ('000'hec)
    Yield (kg/h)
  
  
    Khairpur
    27.53
    68.75
    61
    42.9
    30.1
    35.9
    36.2
    1751
    27
    194.5
    770.6
  
  
    Nawabshah
    26.24
    68.39
    29
    42
    30.4
    36.1
    42.1
    1749
    28
    109.4
    807.2
  
  
    Sanghar
    26.04
    68.95
    25
    40.9
    29.4
    35.1
    45.5
    1753
    24
    116.3
    802.3
  
  
    Mirpur Khas
    25.51
    69.01
    17
    40.8
    28.9
    34.6
    69.6
    1712
    25
    57.7
    753.1
  
  
    Jhang
    31.28
    72.33
    158
    41.6
    30.3
    35.9
    72.2
    1808
    46
    110.9
    481.1
  
  
    Sahiwal
    30.67
    73.11
    152
    41.6
    30.2
    35.9
    92.5
    1812
    44
    150.8
    524.3
  
  
    Multan
    30.15
    71.52
    122
    41.1
    30.6
    35.8
    99.6
    1827
    35
    342.6
    639.6
  
  
    Vehari
    30.04
    72.34
    135
    40.9
    30.7
    36.2
    56.3
    1816
    34
    411.4
    640.5
  
  
    Rajanpur
    29.1
    70.33
    97
    41.8
    30.7
    36.6
    44.2
    1801
    32
    298
    624.1
  
  
    Bahawalpur
    29.35
    71.69
    214
    42.3
    31
    36.6
    50.3
    1807
    29
    257.5
    697.2
  
  
    Rahim Yar khan
    28.42
    70.3
    80
    42.1
    30.4
    36.7
    35.7
    1780
    34
    283.2
    673
  
  
    Bahawal Nagar
    30
    73.24
    163
    40.7
    30.6
    36
    57.7
    1809
    36
    198.1
    618.4
  
  
    Note: Climate variable: Tmax: Maximum Temperature; Tmin: Minimum Temperature; Tave: Average Temperature; Pre: Precipitation; Sun: Sunshine Hour; Hum: Average Air Relative    Humidity.



Table 1: The average cotton index, climate variables of cotton growth season (May-Sep) and site information in Indus River basin Pakistan.

Spatial analysis of change in cotton and climatic variables

In this study, trends were analyzed using the Modified Mann- Kendall (MMK) method [39] based on the original non-parametric Mann-Kendall approach [40,41]. MMK statistic is preferred in investigating the climatic changes and trends because it considers the auto-correlation effect and does not require the dataset to be normally distributed, it has low sensitivity to abrupt break due to inhomogeneous time-series [39]. Huang and Li [33] used the MMK test for demonstrating the trends in climate variables in oases of the arid region of China from 1981-2012 and [42] conducted the rainfall trend analysis for the period 1968-2010 using linear regression analysis and MMK test. The MMK statistic Z_c was applied robustly to test the trend in the time series of climate variables and cotton indices from 1989 to 2018 at 12 sites. The hypothesis that S=0 is accepted using a two-tailed test. The Non-parametric test has more potential than parametric when the probability distribution of data is biased [43]. MMK test is preferred when various sites are tested in a single study [44].

We processed the climate and cotton data (Table 1) in ArcGIS. Kriging geostatistical method was used to generate an estimated surface from a scattered set of points containing different values. It effectively involves an interactive investigation of the spatial behavior of the phenomenon represented by different values and this technique not only has the capability of producing a prediction surface but also provides an accuracy of the prediction. The interpolation method assumes that the distance or direction between sample points reflects a spatial correlation that was used to explain the variation in the surface. This technique was used for visualization, precise understating, thematic mapping, and spatial distribution of change in climate variables, and the cotton index was mapped site-specifically using ArcMap 10.5 software.

Influence of climatic changes on cotton yield

We used the multivariate linear regression (MLR) function, and the Best-Subset Selection (BSS) method to further describe the relationship between crop yield and climatic variables. This MLR function, forced through the origin, was written as follows [45].

ΔY = λiΔX

Where ΔY is the first-order difference of the cotton yield index λ_i shows the standardized regression coefficients, ΔX is the firstorder difference of the key climate variables (i=1, 2, . . ., 6). The standardized coefficients have been used to study the impact of the variables when variables are measured on a different scale [46]. Analyzing the influence and strength of the relationship between the climate variable and the cotton yield in North China from 1981- 2012 [47] were used correlation coefficient and MLR analysis. Their outcomes suggested that warming enhanced the growth of cotton and the progression of the cotton phenomenon, but shortened the period of cotton growth. The environmental factor, including climate and non-climatic factors and location; are the leading causes of inter-annual alterations in crop growth. Thus, it is obligatory to eliminate the confounding effect of non-climatic factors [45,48]. This method reduces the impacts of varieties, management factors, and fertilizers to describe the annual crop growth against climate [38,49]. The regression equations between cotton yield and climate variables (T_max, T_min, T_ave, P_re, S_un and H_um) were selected by the BSS regression method. The best-subset selection method can select the best combination of descriptors from the group of independent variables, as all combinations are examined by [50]. In an attempt to decide on the particular subset to be used, many statisticians consider evaluating all possible regressions and then based on some measured of the subset regression, attempt to reach a decision [51]. In this method, the possible number of the prediction models that would be 2^k, where k is the number of independent variables [52,53]. For measuring the monthly climate variable’s impact on cotton yield, the number of independent variables was thirty. Gorman and Toman proposed that (k¦p) possible regressions of size p subset that will be considered to be good regression when k >10. By using the adjusted coefficient of determination, (R²_adj) model stability was assessed. The model with greater the R²_adj values, has more stability at a significant level [54]. Root Mean Square Error (RMSE) was used to assess the predictive ability of the model. Liu et al. recommended that subset having minimum RMSE and large R²_adj being considered as best. Pearson correlation was established between pairs of the first-order difference in cotton yield indices and climatic variables during the growing season [55,56]. The values of the correlation coefficients essentially represented the climate variables that were meaningful for cotton production at a certain level of significance. Statistical calculations were performed by the XLSTAT software package.

Results

Temporal and crop-related variations and trends

All climate variables and crop-related variables fluctuated within ranges of maximum 39-44 °C, minimum 25-32 °C, average temperature 34-38 °C, precipitation 5-455 mm, sunshine hours 1658-1865 h and air humidity 24-48 % and the variation in cotton cultivated area and yield across the different sites of Indus River basin during the past 30 years can be observed (Figure 2). More fluctions in T_ave, T_min and H_um have been noted in often plases. Especially in Nawabshah, Mirpur, Khairpur, Rajanpur, Bahawalpur, Vehari, Multan. Similarly, Khairpur, Mirpur, Nawabshah, and Sanghar areas have seen more variations in cotton yield. Temperature and air relative humidity at different sits shows a significant variation. Almost all sites recorded a low precipitation trend of less than 100mm as well as some extreme events occur at four main cotton cultivated sites. T_min, T_ave and S_un increased significantly in almost all the study sites (Figure 3). Two sites Sahiwal and Rajanpur shows a significant decrease in precipitation trends, while significant increase trend observed in humidity at Multan, Vehari, Jhang, and Sahiwal which is the main cotton cultivation area. The cotton crop cultivated area decreased significantly at three sites in the north of the study area, which is related to Punjab province. Cotton yield shows a significant increasing trend at all four sites in the south of the study area, which are located in Sindh province.

Figure 2: Box plots of the climatic variables and cotton indices during cotton growth season over 1989 - 2018 at 12 sites BN-Bahawalnagar, BP-Bahawalpur, JHJhang, KP-Khairpur, MPK-Mirpur khas, MT-Multan, NBS-Nawabshah, RYK-Rahim yar khan, RP-Rajanpur, SW-Sahiwal, SH-Sanghar and VH-Vehari in Indus basin River, Pakistan. Horizontal line: median; box-boundaries: 25^th and 75^th percentiles; whiskers: 10^th and 90^th percentiles; circles outliers.

    
    
    Figure 2: Box plots of the climatic variables and cotton indices during cotton growth season over 1989 - 2018 at 12 sites BN-Bahawalnagar, BP-Bahawalpur, JHJhang,
KP-Khairpur, MPK-Mirpur khas, MT-Multan, NBS-Nawabshah, RYK-Rahim yar khan, RP-Rajanpur, SW-Sahiwal, SH-Sanghar and VH-Vehari in Indus basin
River, Pakistan. Horizontal line: median; box-boundaries: 25^th and 75^th percentiles; whiskers: 10^th and 90^th percentiles; circles outliers.

Figure 3: Spatial distributions of trend and significance of the climatic and cotton variables averaged for the cotton growth season for the 12 study locations considered over 1989-2018.

    
    
    Figure 3: Spatial distributions of trend and significance of the climatic and cotton variables averaged for the cotton growth season for the 12 study locations
considered over 1989-2018.

To observe the fluctuations in cotton yield, area, and trends of climate change during crop season in study area during past thirty years, trends lines have been drawn. Cotton crop area and precipitation shown decline while the rest have been increasing. To know more about the trends of change, a fifteen-year trends line has also been drawn in which the minimum temperature, crop area and yield rapidly changing trends have been observed Figure 4.

Figure 4: Linear trend tines of cotton indices and climate variables. Blue lines show trends in variables from 1989-2018 while rend lines from 2004-2018.

    
    
    Figure 4: Linear trend tines of cotton indices and climate variables. Blue lines show trends in variables from 1989-2018 while rend lines from 2004-2018.

Figure 5: Bar chart of climate variables linear regressed on cotton yield (kg/h) blue bar show standardized coefficients, which are statistically significant at the 5% level of significance.

    
    
    Figure 5: Bar chart of climate variables linear regressed on cotton yield (kg/h) blue bar show standardized coefficients, which are statistically significant at the 5%
level of significance.

Selection of the best equation

To describe the selection procedure of the best equation we performed a single variable regression for the study site Sanghar was selected (Table 2). The equation namely, ΔY = -0.63ΔT_ave was found to be the best with the largest R²_adj and the smallest RMSE value among the six equations. Furthermore, the MLR method on the basis of Pearson’s correlation is shown to be a key parameter in the cotton index was performed for cotton yield with the 2, 3 -,… to 6 variables respectively. Finally, equation ΔY = 0.71ΔT_max + 0.24ΔT_min – 0.72ΔT_ave + 0.56ΔP_re + 0.28ΔSun was selected as the best among the all-possible subsets equations for Sanghar. Only the key climate variables that may significantly impact cotton yield were selected through this process, and the other variables were eliminated.

Table 2: Single variable fitted equation for ΔY and their performance at the Sanghar site.




  
    Climate Variable
    Equation
    R2adj
    RMSE
    P-value
  
  
    1
    ΔY = 0.21ΔTmax
    0.01
    327.3
    <0.30
  
  
    2
    ΔY = -0.39ΔTmin
    0.122
    308.4
    <0.05
  
  
    3
    ΔY = -0.63ΔTave
    0.372
    260.8
    <0.0001
  
  
    4
    ΔY = 0.51ΔPre
    0.231
    288.5
    <0.005
  
  
    5
    ΔY = -0.02ΔSun
    -0.035
    334.7
    <0.90
  
  
    6
    ΔY = 0.07ΔHum
    -0.026
    333.2
    <0.7



Table 2: Single variable fitted equation for ΔY and their performance at the
Sanghar site.

Similarly, this procedure was repeated for all 12 study sites for the selection of the best multivariate regression equations for monthly climate variables and their statistical properties are described (Table 3). For viewing the impact of climate variables on cotton yield in the growing season, the best fitted MLR models standardize coefficients were plotted in (Figure 4).

Table 3: For the cotton growing, season (May-September) the best-fitted multivariate linear regression equation and statistical results of all selected sites.




  
    Site Name
    Equation
    R2adj
    RMSE
    P-Value
  
  
    Khairpur
    ΔY = 0.26ΔTmax(May) + 0.19ΔTmax(June) - 0.41ΔTmax(Sep) + 0.33ΔTave(May) –

      0.53ΔSun(Aug) - 0.51ΔHum(May) - 0.57ΔHum(Aug) + 0.20ΔHum(Sep)
    0.89
    88.9
    <0.0001
  
  
    Nawabshah
    ΔY = -0.41ΔTmax(Aug) + 1.27ΔTmax(Sep) - 0.84ΔTave(Sep) + 0.30ΔPre(May) +

      0.32ΔSun(June) - 0.35ΔSun(Sep) - 0.72ΔHum(Sep) + 0.42ΔHum(Sep)
    0.82
    124.8
    <0.0001
  
  
    Sanghar
    ΔY = 0.56ΔTmax(June) - 0.45ΔTave(June) - 0.28ΔTave(July) - 0.38ΔPre(May) +

      0.77ΔPre(Aug) - 0.24ΔHum(May) - 0.31ΔHum(June) - 0.29ΔHum(Aug)
    0.84
    130.2
    <0.0001
  
  
    Mirpur Khas
    ΔY = 0.25ΔTmax(July) + 0.47ΔTmin(May) + 0.46ΔTmin(July) - 0.59ΔTmin(Sep) +

      0.59ΔSun(May) - 0.30ΔSun(Aug) - 0.21ΔHum(Aug)
    0.87
    104
    <0.0001
  
  
    Rahim Yar Khan
    ΔY = -1.04ΔTmin(May) + 1.98ΔTmin(July) - 0.70ΔTave(Aug) - 0.85ΔPre(Aug) +

      0.53ΔSun(July) + 0.82ΔHum(May) + 0.78ΔHum(July)
    0.53
    75.1
    0.003
  
  
    Rajanpur
    ΔY = 0.41ΔTmax(July) + 0.96ΔTmin(Aug) - 0.61ΔTave(Aug) + 0.52ΔSun(Sep) +

      0.41ΔHum(June) + 0.79ΔHum(Sep)
    0.62
    61.1
    <0.0001
  
  
    Bahawalpur
    ΔY = 2.1ΔTmax(May) + 0.54ΔTmax(July) + 2.9ΔTmin(May) - 3.97ΔTave(May) -

      0.46ΔPre(Sep) + 0.84ΔSun(Sep) + 0.54ΔHum(June) + 0.53ΔHum(Aug) + 0.62ΔHum(Sep)
    0.77
    57.1
    <0.0001
  
  
    Bahawalnagar
    ΔY = -1.51ΔTmax(June) - 1.28ΔTmin(June) + 0.57ΔTmin(July) + 1.53ΔTave(June)
    0.63
    77.3
    0.014
  
  
    Multan
    ΔY = -1.1ΔTmax(June) + 0.98ΔTmax(July) + 0.61ΔTave(June) - 0.36ΔSun(June) +

      0.33ΔSun(Sep) - 0.55ΔHum(July) - 0.35ΔHum(Aug) + 0.71ΔHum(Sep)
    0.75
    78.2
    0.001
  
  
    Vehari
    ΔY = 0.85ΔTmin(June) - 1.94ΔTmin(July) + 1.78ΔTave(July) - 0.93ΔHum(Aug) + 0.67ΔHum(Sep)
    0.65
    86.2
    0.002
  
  
    Sahiwal
    ΔY = 0.86ΔTmin(June) - 0.96ΔTmin(July) + 1.16ΔTave(July) + 0.55ΔSun(July) + 0.66ΔHum(July)
    0.34
    123.1
    0.04
  
  
    Jhang
    ΔY = 0.56ΔTmax(July) + 0.49ΔTmin(June) + 0.41ΔPre(Sep) + 0.83ΔSun(July)
    0.47
    92.4
    0.003



Table 3: For the cotton growing, season (May-September) the best-fitted multivariate linear regression equation and statistical results of all selected sites.

These standardized coefficients bar show the direction and strength of the climate variables that significantly impact cotton yield.

Intersection between cotton yield and climate variables

Negative correlation between cotton yield and climate variables has been found in most sites during cotton crop season months. In arid areas, there is positive and negative correlation mix in different months. While in the upper areas where the humidity is high, the negative correlation is more noticeable. In Multan, Vehari, Jhang, and Sahiwal, air humidity, precipitation, and minimum temperature were negatively correlated with cotton yield especially in May and June. While Nawabshah, Sanghar, and Khairpur areas show positive correlation with Tave and Sun in June and July Figure 6.

Download PDF

Citation:Naveed M, He HS, Yasir QM, Du H and Satti Z. Analyzing the Impact of Climate Change on Cotton Yield Using Spatial Analysis and Statistical Modeling in the Indus River Basin, Pakistan. Ann Agric Crop Sci. 2021; 6(5): 1089.

Home

Journal Scope

editorial Board

Instruction for Authors

Submit Your Article

Analyzing the Impact of Climate Change on Cotton Yield Using Spatial Analysis and Statistical Modeling in the Indus River Basin, Pakistan

Abstract

Abbreviations

Introduction

Methods

Study area and datasets

Spatial analysis of change in cotton and climatic variables

Influence of climatic changes on cotton yield

Results

Temporal and crop-related variations and trends

Selection of the best equation

Intersection between cotton yield and climate variables

Discussion

Conclusion

Acknowledgment

References

Site Name	Equation	R2adj	RMSE	P-Value
Khairpur	ΔY = 0.26ΔTmax(May) + 0.19ΔTmax(June) - 0.41ΔTmax(Sep) + 0.33ΔTave(May) – 0.53ΔSun(Aug) - 0.51ΔHum(May) - 0.57ΔHum(Aug) + 0.20ΔHum(Sep)	0.89	88.9	<0.0001
Nawabshah	ΔY = -0.41ΔTmax(Aug) + 1.27ΔTmax(Sep) - 0.84ΔTave(Sep) + 0.30ΔPre(May) + 0.32ΔSun(June) - 0.35ΔSun(Sep) - 0.72ΔHum(Sep) + 0.42ΔHum(Sep)	0.82	124.8	<0.0001
Sanghar	ΔY = 0.56ΔTmax(June) - 0.45ΔTave(June) - 0.28ΔTave(July) - 0.38ΔPre(May) + 0.77ΔPre(Aug) - 0.24ΔHum(May) - 0.31ΔHum(June) - 0.29ΔHum(Aug)	0.84	130.2	<0.0001
Mirpur Khas	ΔY = 0.25ΔTmax(July) + 0.47ΔTmin(May) + 0.46ΔTmin(July) - 0.59ΔTmin(Sep) + 0.59ΔSun(May) - 0.30ΔSun(Aug) - 0.21ΔHum(Aug)	0.87	104	<0.0001
Rahim Yar Khan	ΔY = -1.04ΔTmin(May) + 1.98ΔTmin(July) - 0.70ΔTave(Aug) - 0.85ΔPre(Aug) + 0.53ΔSun(July) + 0.82ΔHum(May) + 0.78ΔHum(July)	0.53	75.1	0.003
Rajanpur	ΔY = 0.41ΔTmax(July) + 0.96ΔTmin(Aug) - 0.61ΔTave(Aug) + 0.52ΔSun(Sep) + 0.41ΔHum(June) + 0.79ΔHum(Sep)	0.62	61.1	<0.0001
Bahawalpur	ΔY = 2.1ΔTmax(May) + 0.54ΔTmax(July) + 2.9ΔTmin(May) - 3.97ΔTave(May) - 0.46ΔPre(Sep) + 0.84ΔSun(Sep) + 0.54ΔHum(June) + 0.53ΔHum(Aug) + 0.62ΔHum(Sep)	0.77	57.1	<0.0001
Bahawalnagar	ΔY = -1.51ΔTmax(June) - 1.28ΔTmin(June) + 0.57ΔTmin(July) + 1.53ΔTave(June)	0.63	77.3	0.014
Multan	ΔY = -1.1ΔTmax(June) + 0.98ΔTmax(July) + 0.61ΔTave(June) - 0.36ΔSun(June) + 0.33ΔSun(Sep) - 0.55ΔHum(July) - 0.35ΔHum(Aug) + 0.71ΔHum(Sep)	0.75	78.2	0.001
Vehari	ΔY = 0.85ΔTmin(June) - 1.94ΔTmin(July) + 1.78ΔTave(July) - 0.93ΔHum(Aug) + 0.67ΔHum(Sep)	0.65	86.2	0.002
Sahiwal	ΔY = 0.86ΔTmin(June) - 0.96ΔTmin(July) + 1.16ΔTave(July) + 0.55ΔSun(July) + 0.66ΔHum(July)	0.34	123.1	0.04
Jhang	ΔY = 0.56ΔTmax(July) + 0.49ΔTmin(June) + 0.41ΔPre(Sep) + 0.83ΔSun(July)	0.47	92.4	0.003