Association of weather variables with yield and yield components of cotton ( Gossypium hirsutun L . ) at reproductive phenophase

Growth and development of cotton is influenced by several environmental factors such as fertility and biology of soil, change in temperature, amount and distribution of rainfall and carbon dioxide concentration which are attributes of climate change. A field experiment was conducted to study the contribution of weather variables for the total variation in yield and yield components during reproductive phenophase of cotton (square initiation to boll opening) during 2013-14 and 2014-15 kharif season. The experiment was set out with three sowing time (24, 26 and 28 standard week) as main plot, three deficit irrigation schedules (0.8 IW/CPE, 0.6 IW/CPE and 0.4 IW/CPE) and rainfed as sub plot in split plot design. The result indicated that significant effect of sowing time on yield and yield components of cotton where maximum number of bolls per plant, boll weight, hundred seed weight, ginning percent, and seed cotton yield attained on early sown cotton. Weather variables showed significant correlation with yield and yield components. Results of regression analysis suggested that weather variables such as maximum temperature, mean temperature, relative humidity I and relative humidity II were found to be influential and accounted for over 90% of total variation in seed cotton yield and yield components during the reproductive phenophase.


INTRODUCTION
Environment for optimum plant growth and development plays a vital role in realizing crop growth and yields.The time of sowings as varied growth condition for various crops differs depending on climate, varieties and method of cultivation (whether rainfed or irrigated).Knowledge on effects of various elements of environment on crop growth, development and yield is important for agronomists and crop production specialists.Cotton experiences temperature fluctuations ranging from 5 to 45°C during the season, which adversely influence *Corresponding author.E-mail: mekonnenadare@yahoo.co.uk.
Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License growth and development (Reddy, 1994).Study reports also indicated that the high temperatures combined with water stress result in boll shed, small boll size and leaf damage (Hake and Silvertooth, 1990).Reddy et al. (1999) reported a significant decrease in boll growth followed by fruit shed within 3 to 5 days after blossom when there was an increase in temperatures over 32ºC.Sustained changes in temperature during the fiber thickening period will also lead to differences in micronaire (Bange, 2010).Quantitative information regarding plant responses to weather change, soil and management condition essential to design crop adaptation mechanisms and improving productivity.
So far, many researches were undertaken under controlled environment conditions and sowing dates at field conditions.Pettigrew (2002) studied cotton response towards early planting over normal planting date.Dong et al. (2006) compared yield, quality and leaf senescence of cotton of late planting production system and normal planting production system.Moreover, enhancing UV-B radiation effects by Kakani et al. (2003) and temperature influences by Bange (2007) were studied on cotton growth and development.However, these studies did not reveal the relationship among the climatic factors under field situations with yield and yield components.Changes in yields attributes to a single factor such as temperature is not possible due to the confounding effects like rainfall, and solar radiation during crop growing period.Furthermore, controlled environment studies often underestimate yield losses from temperature effects at different phenology of a crop that would occur under field conditions (Paulsen, 1994).Furthermore, studies of crop management practices accompanied with weather conditions plays vital role to provide valuable information for producers.Therefore, this study was designed to identify the critical weather parameters affecting the yield and yield components on cotton.

MATERIALS AND METHODS
A field experiment was carried out for two consecutive Kharif seasons (Mansoon) (2013/14 -2014/15) at Professor Jayashankar Telangana State Agricultural University College Farm, Rajendranagar, Hyderabad, India, located at an elevation of 542.6 m above sea level and lies within 17°19'19.64''N latitude and 78°24'29.89''E longitudes.The experiment was designed with three time of sowing at two weeks interval viz., 24 standard week (SW) (D1), 27 SW (D2) and 29 SW (D3) during 2013-14 kharif season and 24 SW (D1), 26 SW (D2) and 28 SW (D3) during 2014-15 kharif season as main plot and three deficit irrigations schedule as 0.4 IW/CPE (I1), 0.6 IW/CPE (I2), 0.8 IW/CPE (I3) and rainfed as subplot treatments.The treatments were laid out in split plot design in three replications.Irrigation water of 50 mm was applied for the respective treatments when the CPE reaches 125.0, 82.3 and 62.3 mm.Cotton seeds (cultivar Neeraja) were dibbled at 60 × 90 cm spacing and thinned at first two leaf stage to maintain the population to one plant per hill.Recommended fertilizer rate of 120: 60: 60 kg per ha of N, P2O5 and K2O were split applied at 30, 60 and 90 days after sowing to all plots.Harvesting was done manually from a net plot area of 6.48 square meter for four successive pickings and expressed in kilo gram per hectare.Daily weather data on maximum temperature, minimum temperature, mean temperature, relative humidity I, relative humidity II, rainfall, sunshine, wind speed and pan evaporation were collected from weather station, Agricultural Research Institute, Rajendranagar, Hyderabad, India.Data on canopy temperature was recorded on daily bases with infrared thermometer and solar radiation with photo-radiometer at reproductive stage.While the ambient temperature (°C) observed during the period was simultaneously recorded with multi-thermometer.
Daily recorded canopy temperatures and Canopy temperature (TC) -Ambient temperature (TA) difference summarized per phonological stages was used to calculate stress degree day and the summarized data during reproductive was used for statistical analysis.Stress degree day (SDD) was calculated using the following formula as suggested by Idso et al. (1977b): Total heat requirement for each developmental stage was estimated by accumulating degree days.The following formula was used to compute accumulated degree days from the starting point to predict when the developmental stage was reached.

( ) ∑( )
Where Tmean = mean temperature of daily maximum and minimum, Tbase = base temperature for cotton (15.5°C), n= number of days between two stages of development.
Mean of weather data of each sowing time that the crop experienced during reproductive stage (73 to 76 mean days after sowing), from square initiation to boll opening stage, were used for correlation regression analysis with yield and yield components of each season mean.After establishing the relationship between weather and yield and yield attributes step down regression analysis (Draper and Smith, 1996) were calculated with SAS software to study the influence of weather parameters on yield and yield attributes.By this analysis, the contribution of respective weather parameters in bringing out the change in yield parameters was known and prediction equation was worked out.

Analysis of variance of yield and yield components
The number of bolls per plant, boll weight, hundred seed weight, ginning percent and seed cotton yield was significantly influenced due to sowing time in both seasons (Table 1).The number of bolls per plant ranged from 16.6 to 27.1, 24.7 to 27.0 and 20.6 to 26.1 during first year (S 1 ), second year (S 2 ) and pooled mean, respectively.Higher number of bolls per plant was recorded at D 2 (26 SW) followed by D 1 (24 SW).Boll weight ranged from 2.9 to 3.7, 3.6 to 4.3 and 3.2 to 4.0 g during first year (S 1 ), second year (S 2 ) and pooled means, respectively.High boll weight was recorded during early sowing times (D 1 followed by D 2 ) in both seasons.The HSW of S 1 , S 2 and pooled means varied from 8.5 to 9.2, 9.3 to 10.5 and 9.0 to 9.8 g, respectively.
Table 1.Seed cotton yield (kg ha -1 ) and its components as influenced by sowing time and deficit irrigation during S1 and S2 kharif seasons.Sowing time one (D 1 ) followed by D 2 resulted in high mean HSW of 9.8 and 9.2 g, respectively when compared to D 3 (9.0g).Different sowing times create natural variation in growing conditions for a crop.Ginning percentage ranged from 29.7 to 32.5 and 31.9 to 34.1% during S 1 and S 2 , respectively.The result indicated that high ginning percent was recorded for early sowing time during S 1 and late sowing time during S 2 .This implied that dry and hot growing conditions showed as positive impact on ginning percent rather than moist and cool conditions.Seed cotton yield ranged from 952 to 1712, 1798 to 2228 and 1375 to 1970 kg ha -1 during first year (S 1 ), second year (S 2 ) and pooled means.Significantly high mean seed cotton yield was recorded due to D 1 (1970 kg ha ).Cotton grown at sowing time three (D 3 ) during S 1 was severely affected by heavy rainfall which resulted in mass of square fall, boll shed, lodging, and water logging so that the seed cotton yield of D 3 of S 1 was quite low when compared with seed cotton yield of S 2 and pooled means.The result evidenced strong influence of growing environment on yield and yield components through sowing date variation.All the results supported early sowing times (26 followed by 24 SW) for increasing growth and development of cotton plant.Early sowing time gets favour in day length to attain normal crop growth and development which escape effect of the cool weather, late rainfall, disease and pest infestation.Similar results were also reported by Dong et al. (2006) who stated significantly low number of bolls in late planting production systems.Reduced boll weight with delayed sowing time was also documented by Ali et al. (2009).Seed weights are components of yield where its contribution strongly influenced by the environment (Johnson et al., 1996).In this connection, significant effect of sowing time on HSW where the higher HSW was produced by early sown maize plants also documented by Dahmardeh (2012).Similarly, Pettigrew (2002) reported that early planted cotton recorded higher lint (about 8%), number of bolls m -2 and lint percentage significantly than did cotton in the normal planting.
Ginning percent at picking two (GINNII) and four (GINNIV) was negatively and significantly correlated with pan evaporation (r = -0.864*and -0.874*), respectively.Ginning percent at picking four (GINNIV) was also showed negative and significant correlation with stress degree day (r = -0.822*).Low lint percent with insufficient carbohydrate production due to high temperature was demonstrated by Oosterhuis (1999) and Onder et al. (2009); they stated that highest number of opened boll and maximum lint percent resulted from plots under stress condition.Negative effects of relative humidity could be associated with low temperature and disease and pest occurrence which could affect yield and yield components (Blanc et al., 2008).

Regression analysis of yield and yield components with weather variables
Weather variables that showed significant influence on yield and yield components during reproductive phenophase was subjected for step down regression analysis and estimation model for yield and yield components is presented in Table 4.
The mean maximum temperature range between 28.99 to 33.19°C accounted for 89 and 91% of total variation in boll weight and hundred seed weight, respectively.The effect of regressor variable was significant and revealed positive regression coefficient.The result also suggested that other yield and yield components was not significantly affected due to change in mean maximum temperature.Similarly, significant effect of maximum air temperature on flower and boll production was documented by Sawan (2014) during the study of climatic variables at pre and post anthesis on boll setting.Similar results were reported by Ratnam et al. (2014) who stated positive and significant correlation of boll weight, bolls per plant, dry matter accumulation and seed cotton yield with total rainfall, minimum and maximum temperature and relative humidity.
A change in mean minimum temperature over a range of 21.7 to 23.7°C accounted for 82 and 89 percent of total variation in boll weight and seed cotton yield, respectively, over different sowing times and seasons.The effect of regressor variable was significant and revealed positive regression coefficient.Similar results were reported by Ratnam et al. (2014).
A change in mean temperature over a range of 25.4 to 28.4°C accounted for 97, 92 and 74% of total variation in boll weight, hundred seed weight and seed cotton yield respectively, over different sowing times and seasons.The effect of regressor variable was significant and revealed positive regression coefficient.Similar results on boll weight was reported by Ratnam et al. (2014).Positive effects of increasing average temperatures, especially at the start and end of the cotton season, on growth, development and ultimately yield were reported by Bange (2007).Benefits of temperature under optimum range were also documented by Sankaranarayanan et al. (2010) and Reddy et al. (1991).
A change in mean maximum relative humidity over a range of 79.0 to 89.1% accounted for 95, 96 and 73% of total variation in boll weight, hundred seed weight and seed cotton yield, respectively over different sowing times and seasons.The effect of regressor variable was significant and revealed negative regression coefficient suggesting the reduction in yield and yield components during increase in maximum relative humidity.Similar research reports were documented by Oosterhuis (1999), Oosterhuis (1997) and Ratnam et al. (2014).
A change in mean minimum relative humidity over a range of 55.3 to 74.2% accounted for 94 and 96% of total variation in boll weight and hundred seed weight, respectively over different sowing times and seasons.The effect of regressor variable was significant and revealed negative regression coefficient.The present result did not agree with the reports of Ratnam et al. (2014).However, increase in minimum relative humidity could favour the incidence of pest and disease (Sharma and Razdan, 2013) which could adversely affect yield and yield components, so that inverse association could be observed.
A change in total rainfall over a range of 118.0 to 387.2 mm accounted for 68 and 74% of total variation in boll weight and hundred seed weight, respectively over different sowing times and seasons.The effect of regressor variable was significant and revealed negative regression coefficient.A large the amount of rainfall could result in over saturation and water logging which had adverse effect on cotton growth and development.Similarly, in relation to the amount of rainfall, Ogbuene (2010) reported that rice yield was negatively and significantly correlated with rainy days.A change in mean wind speed over a range of 6.1 to 12.3 km h -1 accounted for 88, 78 and 86% of total variation in boll weight, hundred seed weight and seed cotton yield, respectively over different sowing times and seasons.The effect of regressor variable was significant and revealed positive regression coefficient which could be expected in relation to turbulence effect of wind on atmospheric CO 2 , moisture and leaf temperature favouring rate of photosynthesis in cotton.However, Barker et al. (1985) reported plants grown under partial shelter (reduced wind speed) condition resulted in increased plant height, earlier squaring, earlier boll set and more bolls and biomass.
A change in total pan evaporation over a range of 159.3 to 301.0 mm accounted for 75 and 76% of total variation in ginning percent at picking two and four, respectively over different sowing times and seasons.The effect of regressor variable was significant and recorded positive regression coefficient with ginning percent at picking two and negative with ginning percent at picking four.Similarly, a change in stress degree day over a range of 35.5 to 113.6°C accounted for 68% of total variation in ginning percent at picking four over different sowing times and seasons.The effect of regressor variable was significant and revealed negative regression coefficient.Crop growth and development could be benefited from increasing pan evaporation provided that the soil moisture content maintained during critical crop growth periods, however, excess loss of moisture could result in desiccation which adversely affected ginning percent.Similar findings in relation to deficit irrigation were documented by Onder et al. (2009) who reported high ginning percent under stress conditions.

Conclusion
Yield and yield components of cotton was significantly influenced due to sowing time and deficit irrigation.The result suggested that the different sowing times and deficit irrigation created natural variation in growing conditions for cotton crop were attributed with varied climatic elements affected yield and its components.Weather variables such as maximum temperature, mean temperature, relative humidity I and relative humidity II were found to be influential and accounted for over 90% of total variation in seed cotton yield and yield components during reproductive phenophase.

Table 2 .
Mean of weather variables prevailed for sowing times during reproductive phenophase of the study seasons.

Table 3 .
Correlation coefficients (r) of weather variables with yield and yield components during flowering.

Table 4 .
Estimation of yield and yield components by linear regression functions of weather variables at square initiation and first flower phenophases.