Acknowledgements

I would like to acknowledge and give my thanks to Dr. Anoop Kumar Barooah (Former Director, TTRI, TRA) and Dr. Rupanjali Deb Baruah (Scientist, Climate and GIS Lab, TTRI, TRA) for their continuous support and guidance throughout writing this paper.

Abstract

Climate change impacts on tea ecosystem are expected to be highly significant in coming years, evidence of which has already started appearing. There are already proven changes in North East Indian climate which have started affecting tea productivity in one way or the other. Perceptible changes in rainfall and temperature have been noticed; increased weather extremes have complicated the issue further. Intermittent harvest losses, loss of biodiversity in tea eco-system, droughts and floods alternating, increased landslides and erosion in sloppy areas are some of the many problems triggered by climate change. These evidences of disturbance in tea ecosystem may question the sustainability of tea ecosystem in the long run. Forecasting crop yield well before harvest is crucial especially in regions characterized by climatic uncertainties. This enables planners and decision makers to predict how much to import in case of shortfall or to export in case of surplus. Therefore, in this paper an attempt has been made to study the effect of various climatic variables on yield of tea crop. While doing so correlation studies between different climatic variables have been performed as well as dependence of yield on specific climate variables has been identified by carrying out multiple regression analysis. Accordingly, best fitting crop yield prediction equations were derived and applies to find out how well predicted and observed production matched.

Introduction

Tea crops produced in northeast India are of major importance for the regional economy. India is the second largest producer and exporter of tea (Tea Board of India, 2014) with the north- eastern state of Assam being a key producing region. The predominant tea variety produced in Assam is Assam-type var. Assamica (De Costa et al., 2007). On its own, Assam contributes around 17% of world tea production and annually produces more than 50% of India’s tea (Dikshit &

Dikshit, 2014). Tea also plays a pivotal role in supporting the livelihoods of approximately 1.2 million laborers in Assam (Dikshit & Dikshit, 2014). The tea bush grows in a specific climatic niche, thus making the industry vulnerable to climatic variability and climate change. Given the importance of the tea crop to the economy in northeast India, and as an employer, understanding its sensitivity to climatic variation is of paramount importance. Thus, in this study, an attempt has been made to study the effect of various climatic variables on yield of tea crop. While doing so correlation studies between different climatic variables were carried out as well as dependence of yield on specific climate variables were identified by carrying out multiple regression analysis.

Literature review

Usually, laboratory or experimental field studies have been used to assess the impact of climatic variation on tea productivity and growth (Carr, 1972; Hadfield, 1975; De Costa et al., 2007). Recent studies suggest that increased water availability increases tea plant growth and growth of new leaves on tea bushes (Ahmed, Orians, Griffin et al (2013). These same authors also identify that climatic variation can influence the quality of tea production (Ahmed et al., 2013, 2014; Carr, 1972; Hadfield,1975; De Costa et al., 2007). While experimental studies are useful to identify the effect of specific mechanisms through which climatic variation can impact tea productivity, they do not monitor the impact of climatic variation on tea productivity as it occurs in farmers’ fields. Other approaches to assess climate impacts on crop productivity utilize crop simulation models; however, crop simulation model structure may not allow identification of the impact of climate extremes and such models require large amounts of data for calibration (White, Hoogenboom, Kimball, & Wall, 2011). To address these shortcomings, the climate-impacts literature has undertaken econometric analysis of weather on realized crop yields. Lobell and Burke (2010) and Hsiang (2016) provide overviews of the econometric methods used. Typically,

regression models are estimated using panel datasets of crop yield (Burke &Emerick, 2016; Lobell, B€anziger, Magorokosho, & Vivek, 2011; Schlenker & Lobell, 2010; Schlenker & Roberts, 2009). For example, Schlenker and Roberts (2009) and Burke and Emerick (2016) identify nonlinear heat impacts on maize and soy yield using county-level data in the USA. Burke and Emerick (2016) also use a long-differences model to capture potential effects of adaptation, and show little evidence of adaptation over the past 50 years. District-level panel data of drought intensity, rice yield, and irrigation from India identifies drought impacts on rice yield, but also decreasing sensitivity of rice yields to drought over time (Birthal, Negi, Khan, & Agarwal, 2015). Regression models trained with panel data have also been applied to experimental station crop yields for maize (Lobell et al., 2011) and remote sensing data for wheat (Duncan, Dash, & Atkinson, 2014; Lobell, Sibley, & Ivan Ortiz-Monasterio, 2012) to identify climate impacts on yield. A recent analysis has used panel data of tea production in China to identify the effect of monsoon dynamics and weather on tea production (Boehm et al., 2016). An increase in retreat date of the monsoon, and an increase in monsoon precipitation is associated with a decrease in tea yield (Boehm et al., 2016).

Despite the regional importance of the tea crop in Assam, and its potential vulnerability to climate change, limited assessment of the effect of climatic variability on tea productivity in operational plantations has been undertaken. Thus, we have little idea of the extent to which climatic variability is impacting tea productivity in Assam or the extent of the resilience or sensitivity of the real-world tea production systems to climate change (based on the assumption that yield response to contemporary weather is a proxy for response to climate change (Hsiang, 2016)). Further, if climatic variability is affecting tea yield, we do not know which climatic variables are of importance and should be targeted with adaptation measures. To address this gap, a unique dataset of tea yield and climatic parameters were collected from a garden location in

South bank region of Assam for 10 years. There are several climatic factors that influence the production of tea in Assam. The region has a humid, subtropical climate with high levels of rainfall, which is ideal for the cultivation of tea. The warm, humid climate of Assam allows for a long growing season for the tea plants, which helps to produce high-quality tea leaves. One of the most important climatic factors affecting tea production in Assam is the monsoon season. The monsoon season in Assam begins in June and lasts until September, and it brings heavy rainfall to the region. The monsoon season is crucial for the growth of the tea plants as it provides the plants with the necessary water and nutrients they need to thrive. However, excessive rainfall during the monsoon season can also lead to problems, such as flooding and landslides, which can damage the tea plants and adversely affect the quality of the tea leaves.

Another important climatic factor affecting tea production in Assam is temperature. The temperature in Assam is generally warm and humid throughout the year, but it can vary significantly between the different seasons. The temperature during the summer months can reach up to 40°C, while the winter months can be much cooler, with temperatures dropping to around 10°C. These variations in temperature can affect the growth and development of the tea plants and the quality of the tea leaves.

In addition to climatic factors, other environmental factors can also impact tea production in Assam. One such factor is soil quality. The soil in Assam is generally rich in nutrients and well- draining, which is ideal for the cultivation of tea. However, soil erosion and overuse of pesticides and fertilizers can lead to soil degradation, which can have negative effects on the tea plants and the quality of the tea leaves.

Materials and methods

Historical archived data available with the Climate Research Centre of Tocklai Tea Research Institute, Jorhat, Assam was used for the study. The data set included data of rainfall, maximum and minimum temperature, and section – wise yield data ranging for the period 2011 – 2020. In the Climate Research Centre of Tocklai Tea Research Institute rain gauges with varying diameters—6.28 inches and 8 inches (Self Recording Rain Gauges)—are used to measure rainfall. Symon’s design Rain gauge is its name. The cylindrical body of the funnel contains a jar where the rainwater is collected and the amount of rain is measured using a graduated measuring glass.

Figure 1

Symon’s pattern Rain Gauge

Stevenson Screen is used to measure the maximum and minimum temperature. It is a wooden box with two roofs and twin louvres. The thermometers’ bulbs should have unrestricted access to air, but they shouldn’t be exposed to sunlight or precipitation. There are four thermometers in total:

1. Dry bulb thermometer (use mercury bulb)
2. Wet bulb thermometer (use mercury bulb)
3. Max. temperature (use mercury bulb)
4. Min temperature (use alcohol bulb)

Figure 2

Stevenson Screen

The trends analysis was performed for the two climatic factors i.e., rainfall, maximum and minimum temperature with respect to the annual yield. The total rainfall, average maximum and minimum temperature and the average yield was calculated using formulas on excel and the necessary graphs were plotted. Accordingly, the trend lines were drawn on the line graphs that were obtained to clearly identify an increase or decrease in the weather parameters. Similar graph and trend line was drawn for the annual yield.

The statistical method that was used for determining the trend between the above-mentioned factors was Multiple Linear Regression (MLR). The annual yield of tea was plotted against each of the climatic variables and a line of best fit was drawn to get a basic idea of the trend. A linear equation y = ax + b was obtained by the regression graph, where a is the slope of the graph, b is the y intercept of the graph, y is the annual yield of the tea and x is the climatic factor being

analysed. The linear trend value represented by the slope ‘a’ of the simple least square regression line provides the rate of increase or decrease in the annual yield with the climatic factors.

Results and Discussion

Rainfall trends

The archived data of annual rainfall for Jorhat was used for the study. According to the graph in Fig 4, the positive trend line for the annual rainfall depicts a slight increase in rainfall over the 10 years. However, in general, it has been observed that rainfall for most tea growing regions in Assam is decreasing and for Jorhat (South bank), Assam, alone the rainfall has decreased by more than 200 mm. There are seasonal changes with low and high rainfall occurring, depicted by the changes in peaks in the graph, making some years wet and some years droughty, but the frequency of extreme rainfall events has grown in the previous around 30 years, with more droughts and floods appearing than in the past (data not shown). When the data set was bifurcated for two different seasons of the year as shown in Fig.3, it was observed that the rainfall was very high for the productive season and quite low for the dormant periods. The total rainfall for the productive season varied between 1009.3 mm to 3948.4 mm while for the dormant season under the ten years of study, it varied between 132.7 mm to 683.6 mm.

Figure 3

Figure 4

However, the increase in rainfall over the last ten years is a positive sign for growth of tea in Assam as it helps rainfall direct towards the optimal amount required by the tea leaves.

Minimum Temperature

The archived data of annual rainfall for Jorhat was used for the study. A statistical average trend line was plotted with R2 values to visualize the long-term trends. The minimum temperature (Tmin) values at Jorhat (South bank) Assam shows a decreasing trend of 0.12 mm each year. Most

people agree that tea plants thrive and produce at their best between 13 and 30⁰C (Barua, 1989). But historical data going back more than 90 years indicates that the number of days with temperatures above 30°C and 35°C has been rising for the past 30 years. However, tea continues to thrive in these circumstances, suggesting that the plant has some sort of self-adaptation to escalating temperature extremes. To evaluate cultivars for temperature resilience, controlled condition tests must be conducted at high (and low) temperatures. The extent to which tea can adapt to a rising temperature, though, may still be in doubt.

Figure 5

During the productive season Tmin values varies between 18.9⁰C – 25.0⁰C which lies well within the optimal range. During the dormant season, Tmin values varies between 9.1⁰C – 15.6⁰C which is below the optimal range. During the entire year Tmin values were in general lower for North Bank (data not shown). The temperature analysis shows that, on average, the tea growth temperature requirement is still well above the limiting temperature in the productive season

leading to high yield in tea crops; however, during the dormant season, the continuous fall in averages Tmin is a matter of concern and would require additional spatiotemporal research in this area.

Figure 6

Maximum Temperature

The average maximum temperature (Tmax) datasets were plotted for different time periods for the meteorological station Jorhat located in upper bank of Assam. In general, Tmax shows an increasing trend of 0.017 per annum with an R2    value of 0.033, where the recordings are fairly recent (only for last 10 years). The long-term data for other tea growing regions also demonstrate an increasing trend for maximum temperature which is possibly due to the effects of climate change and global warming. The yearly averages for maximum temperature were mostly above 30⁰C and when the data was bifurcated into dormant and active seasons, it was noticed that the Tmax values for dormant period was between 23.0⁰C – 27.8⁰C whereas for active season, it was between 28.5⁰C – 32.8⁰C. Hence, the results reflect that the average values for maximum temperature have already crossed the upper limit of 13-30⁰C for good growth.

Figure 7

Maximum Temperature for different months year wise

The increasing maximum temperature is a matter of concern for tea gardens in Assam as it is leading to degradation in the tea crops. High temperatures can cause stress on the tea plants, leading to reduced growth and yield.

In addition to this, it can also cause changes in the composition of the tea leaves, resulting in a lower quality product. For example, high temperatures can lead to an increase in the levels of caffeine and tannins in the leaves, which can affect the taste and aroma of the tea. The risk of pests and diseases due to increasing maximum temperatures can further reduce crop yields. Pests such as tea mosquito bug and tea thrips are known to thrive in high temperature conditions which can possibly hamper the growth of tea leaves adversely. The most significant impact of increasing maximum temperature has been the witting of leaves and negative impacts on the photosynthesis process. It has been reported earlier that net photosynthesis increased as the temperature of the assimilating leaves was increased, but beyond 35⁰C, there was a sharp decline and no photosynthesis was observed beyond 39⁰C.

Figure 8

Average Maximum Temperature year wise

Several measures have to be taken in order to mitigate the effects of increasing maximum temperatures. Irrigation has to be a more continuous process and needs to be done more frequently as providing adequate supply of water can reduce the stress caused by high temperatures. Providing shade for the tea plants can help to reduce the amount of direct sunlight they receive, reducing the stress caused by high temperatures. This can be done by using shading nets or by planting trees around the tea fields. Lastly, Covering the soil around the tea plants with mulch can help to reduce evaporation and maintain soil moisture, which can be beneficial during hot weather. Using the multiple regression and standard deviation feature in excel, the equation between the yield and the climatic factors was calculated. The equation derived is:

Yield (Y) = 71575.62 + 4873.21*Rainfall – 58704.93*Max Temperature – 2511.83 * Minimum Temperature

Table 1

Regression Statistics and Errors

Table 2

Regression and Residual Values

Table 3

Regression Analysis with relation to the different climatic factors

Table 4

Yearly data on yield, Rainfall, maximum temperature and minimum temperature

The theoretical yield for each year was calculated using the formula above and then the percentage error was also calculated. By observing the values we obtained, we can clearly infer that between the years 2011 and 2012, the yield had a negative impact since the experimental yield was less than the theoretical yield depicted by the negative sign of percentage error.

Table 5

Actual yield, predicted yield and the percentage error

However, between 2013 and 2016, the growth of tea crops in Jorhat had shown a positive sign because the experimental yield was greater than the theoretical yield, elucidated by the positive sign of the percentage error. Nevertheless, after 2016 tea crops have been in decline due to large negative difference between the experimental and the theoretical yield. We can clearly deduce that in Assam, a significant tea-growing region in North East India, the decrease in rainfall and rise in temperature (both Tmax and Tmin) are of concern for tea growth and production.

Figure 9

Graphical representation of trend in actual and predicted yield year wise

Conclusion

The equations derived for yield prediction can prove to be very useful for the tea industry given the nuances of climate change. The tea industry can get an assumption on how much yield they should expect for a certain year, assuming they have the necessary values for the climatic factors such as rainfall, maximum temperature and minimum temperature. This will allow the tea gardens to set a goal for themselves and work accordingly so that they can achieve an experimental yield which is higher than the expected theoretical yield. Accurate forecasts of these climatic parameters would result in accurate production forecasts in the future. Hence this model will be strong supportive tool for the tea industry in making best decisions for management well in advance in order to achieve maximum returns from the plantations.