Changes in Rainfall and Temperature and Its Impact on Crop Production in Moyamba District, Southern Sierra Leone

Rainfall and temperature are the important variables that are often used to trace climate variability and change. A Perception study and analysis of climatic data were conducted to assess the changes in rainfall and temperature and their impact on crop production in Moyamba district, Sierra Leone. For the perception study, 400 farmers were randomly selected from Farmer-Based Organizations (FBOs) in 4 chiefdoms and 30 Agricultural Extension Workers (AWEs) in the Moyamba district were purposely selected as respondents. Descriptive statistics and Kendall’s test of concordance was used to analyze the data collected from the farmers and AEWs. Data for the analysis of variability and trends of rainfall and temperature from 1991 to 2020 were obtained from the Sierra Leone Meteorological Agency and Njala University and grouped into monthly, seasonal and annual time series. Regression analyses were used to determine the statistical values and trend lines for the seasonal and annual time series data. The Mann-Kendall test and Sen’s Slope Estimator were used to analyze the significance and magnitude of the trends respectively. The results of both studies show evidence of climate change in the Moyamba district. A substantial number of farmers and AEWs perceived a decrease in the annual rainfall amount, length of the rainy season, a late start and end of the rainy season, an increase in the temperature during the day and night, and a shortened harmattan period over the last 30 years. Analysis of the meteorological data shows evidence of variability in the seasonal


Introduction
Climate change is among the major challenges that need urgent attention in the 21 st century. Agriculture is among the largest sectors emitting greenhouse gas (GHGs) in the atmosphere [1]. Rainfall and temperature are the two most important variables in the field of climate sciences and hydrology which are frequently used to trace the extent and magnitude of climate change and variability [2], hence changes in their pattern can affect human health, ecosystems, plants, and animals [3]. Precipitation is a vital part of the hydrologic cycle and changes in its pattern would directly influence the water resources of a given area, thus altering the spatial and temporal distribution of runoff, soil moisture, and groundwater reserve [4]. According to [5], examining the trends of rainfall and temperature is important to understand climate variability and change because both are highly variable spatially and temporarily at different local, regional, and global scales. The local climate in Sierra Leone is tropical with two distinct seasons. The rainy season runs from May to October with an average temperature of 22˚C -25˚C whilst the dry season is from November to April with temperatures reaching 25˚C -27˚C, accompanied by dry, cool winds from the Sahara Desert. The rainy season is largely controlled by the movement of the tropical rain belt, Inter-Tropical Convergence Zone (ITCZ), which oscillates between the northern and southern tropics over the course of a year [6].
In Sierra Leone, agricultural production and productivity are dependent on rainfed agriculture. Agriculture is central to food security and rainfall trends and variability are the frequently mentioned factors contributing to food insecurity [7]. Since 1960, there has been evidence of increasing temperature, decreased and unpredictable rainfall, and substantial climatic hazards such as floods, strong winds, and thunderstorms in Sierra Leone [8]. The decreased and unpredictable rainfall has often resulted in seasonal droughts [9] causing the poor establishment of seedlings and wilting of crops at the start and end of the rainy season respectively thus contributing to low crop yields in the country. The increasing temperature has contributed to the insurgence of new pests and diseases in the country [10]. These have threatened the fragile agricultural system [8] posing serious challenges in the struggle to achieve sustainable food production.
Other studies have shown that climate change will affect all four dimensions of food security, namely food availability, access to food, stability of food supplies, and food utilization [8]. For farmers to adapt effectively to climate change, they must have correct perceptions about the state of the climate and possible future trends. In practice, farmers take decisions in the context of their environment, and differences may exist between perceived and real environments [9].
In the Moyamba district, farming calendars are closely associated with the timing of the local rainfall, therefore improved forecasts of the onset and termination of the rainy season would be of great benefit to smallholder farmers. Regarding this, trend analysis of climatic variables has received a great deal of consideration from scholars recently. The characterization of the intra-and inter-annual trend of meteorological variables in the context of a changing climate is vital for assessing climate-induced changes that have already occurred and this will be further helpful to make predictions for better preparedness to develop feasible adaptation and mitigation strategies in agriculture [11]. In addition, there is a need to assess the perception and knowledge of farmers and agricultural extension workers on climate change and how it affects crop productivity in the Moyamba district. Several studies on farmers' perception of climate change and its impact on crop production and analysis of rainfall and temperature time series data at different periods of records have been conducted throughout the world [12] [13]. Few trend analysis studies on rainfall have been conducted in Sierra Leone [9] [14]. Studies to detect changes in the variability and trends of rainfall and temperature in the last climatological period (1991 to 2020) alongside the perception and knowledge of farmers and agricultural extension workers deserve urgent attention in light of the impact of climate change on crop production in Moyamba district. Therefore, this study was designed to assess the changes in rainfall and temperature and their impact on crop production in the Moyamba district. The specific objectives were to: 1) Assess the perception of farmers and agricultural extension workers versus meteorological data on changes in rainfall and temperature patterns in the Moyamba district.
2) Assess the perception of farmers on the impact of climate change on crop production and their strategies and needs for climate change adaptation and mitigation in the Moyamba district.

Description of the Study Area
The studies were conducted in the Moyamba District. Moyamba district is located in the Southern Province and borders the Atlantic Ocean in the west, Port Loko district and Tonkolili district to the north, Bo district to the east, and Bonthe district to the south. The district is the largest in the Southern Province by geographical area and comprise of fourteen chiefdoms ( Figure 1).
Agriculture remains the mainstay of the district residents and the largest sector of the economy in the district, providing livelihoods for over 71% of the population. Crops grown in the district include oil palm, cereals (maize, rice, sorghum, and millet) starch food crops (yam, cassava, sweet potato, and cocoa yam), and grain legumes (groundnut and cowpea). In addition, cashew, pepper, ginger, pineapple, and sugarcane are popular farm products in the district. Despite the abundance of land and water resources, the majority of the farmers have smallholdings of 0.5 to 2 hectares, operating as basic subsistence food production units.

Field Survey on Farmers' and AEWs' Perception and Knowledge of Climate Change and Its Impact on Crop Production
The study was conducted in the Kori, Bagruwa, Fakunya, and Kaiyamba chiefdoms in Moyamba District (Figure 1). Kori and Bagruwa are the chiefdoms implementing a 20-month (March 2021 to November 2022) project titled where: n is the sample size, Z is the value for the selected alpha level, (1.96 at 95 percent confidence level). p is the estimated proportion of an attribute that is present in the population (0.5). q is 1 − p (0.5). (p) (q) was the estimate of variance. d is the acceptable margin of error for the proportion being estimated (0.05).
Data collection.
Desk reviews were used to collect secondary data. Computer-Assisted Personal Interview (CAPI) was used to collect primary data from the sampled farmers and AEWs using a structured questionnaire. Digital forms of the questionnaire were prepared in CSPro 7.5 and were uploaded on the CSEntry app on android tablets for data collection.

Data analysis
Statistical Package for Social Sciences (SPSS) software was used to analyze the data. Descriptive statistics (frequencies and means) were used to analyze the quantitative variables collected from the personal interviews. Kendall's test of concordance was used to determine the degree of agreement of the farmers and MAF AEWs in ranking their needs for climate change adaptation and mitigation. Microsoft Excel 2019 was used to prepare the tables and figures after analysis.

Analysis of Trends in the Monthly, Seasonal and Annual Rainfall and Temperature Patterns
Acquisition of rainfall and temperature data For a variety of reasons, reliable long-term rainfall and temperature records of and Njala University (biased corrected historical rainfall and temperature time-series data which was generated from the Soil & Water Assessment Tool Climate Change Model) [16]. Quality control of the available time-series data was done to identify errors in data sets that may introduce non-climatic shifts in climate time series which have the potential to affect the quality of data and the results of the climatic trends. Correlation analysis was performed on the time series to check for a statistically significant correlation between the annual time series data before selecting the data source for combination.

Data analysis
The rainfall and mean temperature data were grouped into monthly, seasonal, and annual time series before analysis. The seasonal time series data were grouped into two seasons, namely the dry season (November to April) and the rainy season (May to October).

Linear relationship between year and climatic variables
Regression analysis was used to determine the linear relationship between where x is the independent variable and it is plotted along the x-axis, y is the dependent variable and it is plotted along the y-axis, b is the slope line and a is the intercept (value of y when x = 0). Significance and magnitude of the trend The Mann-Kendall test was used to determine the significance and magnitude of the rainfall and temperature trends for the different time series using the Data Analysis Tool of Real Statistic (XRealStats). A trend is a significant change over time exhibited by a random variable, detectable by statistical parametric and nonparametric procedures. The Mann-Kendall test is a non-parametric approach that has been widely used for the detection of trends in different fields of research including hydrology and climatology [17]. Each data value is compared to all subsequent data values. The initial value, S, is assumed to be 0 (no trend).
If a data value from a later period is higher than a data value from an earlier period, S is incremented by 1. If the data value from a later period is lower than a data value from an earlier period, S is decremented by 1. The net result of all such increments and decrements yields the final value of S. The Mann-Kendall statistic (S) was calculated as follows: where n is the number of data points, x j and x i are the data value in time series i and j (j > i), respectively and sgn(x j − x i ) is the sign function and is calculated as: A very positive value of S is an indicator of an increasing trend, and a negative value indicates a decreasing trend.
Since the sample size was large (N > 10), the test was conducted using a normal distribution, with the mean and the variance calculated as follows: where N is the number of tied (zero difference between compared values) groups, t k is the number of data points in the kth tied group.
The standard normal deviate (Z-statistics) was then computed according to the formula [18] as follows: The presence of a statistical trend was evaluated using the Z value. A positive and negative value of the Z statistics indicates an upward and downward trend respectively. The Z statistic has a normal distribution to test for either upward or downward trend at the α level of significance (usually 5% with Z 0.025 = 1.96), a Z value that is greater than 1.96 indicates a significant trend in the time series data.
The probability value (p-value) from the two-tailed test using the Z value also can be used to test the significant trend. When the p-value is greater than α (0.05), it shows that there is no trend in the time series data.
Sen's slope estimator (β) was used to predict the magnitude of the trend. It is a non-parametric and robust estimate for quantifying monotonic trends in hydrologic time series [19]. A positive value of β connotes an "upward trend"; and a negative value for a "downward trend" [19]. The Sen's slope estimator was calculated as follows: where β is the slope between data points x i and x j , x i is the data measurement at the time i, x j is the data measurement at time j; and j is the time after time i. Table 1  Parallel studies have also reported that the respondents interviewed are middle-aged and old farmers [21]. The average household size of the farmers was 8 with more females than males except in the Bagruwa chiefdom. Ninety-two percent (92%) of the farmers depend on agriculture-related activities as their main source of household income. Sixty-five percent (65%) of the farmers owned the land they cultivated indicating access to land for farming is not a major challenge among farmers in the study area. Moreover, 85%, 77%, and 56% of farmers cultivate rice, cassava, and groundnut as their major crops.

Perception of farmers and AEWS of MAF on changes in rainfall
The results of the perception of farmers and AEWs showed that 92.8% of farmers and 100.0% of AEWs interviewed professed changes in rainfall in the Moyamba district over the last 30 years. Figure 2 presents the perception of Table 1   farmers and AEWS on the annual rainfall amount and length of the rainy season. Among the farmers who perceived changes in rainfall, 44.3% of them perceived a decrease in the annual rainfall amount, 28.8% perceived an increase, 26.1% perceived an unpredictable rainfall whilst 0.8% observed no change in the annual rainfall amount. Considering the length of the rainy season, 46.1% of the farmers perceived a decrease in the length of the rainy season, 40.7% perceived unpredictable, 12.7% perceived an increase, and 0.5% perceived no change in the length of the rainy season. In contrast, the majority (70.0%) of the AEWs interviewed perceived a decrease in the annual rainfall amount and a decrease in the length of the rainy season during the last 30 years. Figure 3 presents the results of the perception of farmers and AEWs at the start and end of the rainy season. The results show that 43.9% of the farmers perceived a late, 29.9% perceived an unpredictable, 25.6% perceived an early, and 0.5% perceived no changes in rainfall at the start of the rainy season, whilst 46.6% perceived a late, 39.4% perceived an unpredictable, 13.5% perceived an early, and 0.3% perceived no changes in rainfall at the end of the rainy season. In contrast with the perception of farmers on changes in rainfall, the majority 56.7% of the AEWs perceived a late start (56.7%) and an early end (60.0%) of the rainy season.
Rainfall distribution trends from 1991 to 2020 in Moyamba district Linear relationship between year and rainfall Figure 4 presents the results of the linear relationship between year and rainfall distribution in the Moyamba district from 1991 to 2020. The results show variability in rainfall distribution and decreasing trend lines for the rainy season and annual rainfall from 1991 to 2020 in the Moyamba district. The R-squared (R 2 ) statistic of the regression models accounts for 0.27%, 2.29%, and 1.94% of the variation in the dry season, rainy season, and annual rainfall distribution respectively. The low R 2 values indicate a very weak relationship between rainfall and year meaning the regression model accounts for less of the variance in the rainfall distribution and that the data points of the rainfall distribution are widely dispersed around the regression line.
Significance and magnitude of the trend Table 3 presents the results of the Mann-Kendall test for the monthly, seasonal and annual time series for rainfall in the Moyamba district from 1991 to 2020. At a 95% confidence limit, the results showed positive values of the MK-Stat and Z-stat for the months of June, October, and November and the dry season time series which indicates an upward trend in the rainfall. In contrast, the values of the MK-Stat and Z-stat for the other months, rainy season, and annual time series were negative which indicates a downward trend in rainfall for the remaining months, the rainy season, and seasonal and annual time series. However, the upward and downward trends in rainfall for the monthly, seasonal and annual time series were not significant (p > 0.05). The results of the Sen's slope estimator showed that the magnitude of the rainfall trend in May which is the usual month for the start of the rainy season decreased by 29.87% indicating the occurrence of few and intermittent rainfall events which may cause seasonal droughts at the start of the planting season in May. Consequently, the planting  time for crops that are susceptible to drought spells at the early vegetative growth stage has gradually been shifted to June to ensure the good establishment of crops in the field [9]. In Sierra Leone, the planting period in the first cropping season is May-June. The amount and frequency of rainfall during this period are critical for farmers, especially if they want to plant early at the start of the rainy season. Due to the erratic rainfall during this period, farmers are faced with the risk of a dry period which may cause poor germination and dying of seedlings. Therefore, farmers need to plant at different dates as a coping strategy to adapt to the erratic pattern of rainfall [22]. The magnitude of the rainfall trend for the months of October which marks the end of the rainy season and November which marks the start of the dry season increased by 67.59% and 89.11% respectively. The result shows that the occurrence and intensity of rainfall events during these months were relatively high and may cause a reduction in the quality of the harvest for crops that require relatively drier periods at maturity. Thus, the start and end of the rainy season have gradually shifted to June and November. These results are in agreement with the findings of the perception of farmers and AEWs who considerably observed a late and unpredictable start of the rainy season and a late and unpredictable end of the rainy season. Even though the length of the rainy season has not changed which contradicts the perception of most of the farmers and AEWs, the magnitude of the rainfall trend for the rainy season decreased by 212.74% indicating lower intensity and rainfall events which could be attributed to the decrease in rainfall trends for the months of May, July, August, and September. If the identified trends continue and amplify, then cultivating twice in the rainy season a 90 days crop might become risky due to possible reduced rainfall and dry spells in the mentioned months. Consequently, only crops with maturity periods which are lesser than 3 months can be cultivated twice during the rainy season to avoid seasonal droughts. The magnitude of the annual rainfall trend decreased by 128.53% indicating the occurrence of seasonal drought spells during the cropping season and drier years in the coming years. This result agrees with the findings of the perception of farmers and AEWs which shows a considerable number of farmers and AEWs observed a decrease in the rainfall amounts during the rainy season. Previous studies in Sierra Leone have shown a decline in early seasonal rainfall and total rainfall amount in the country [9].

Perception of farmers and AEWs on changes in temperature
The results of the study show that 94.0% and 100.0% of the farmers and AEWs interviewed respectively perceived changes in the local temperature in the Moyamba district over the last 30 years. Figure 5 presents the results of the perception of farmers and AEWs on the changes in the local temperature during the day and night. Among those who perceived changes in the local temperature 49.7% of them perceived an increase, 30.3% perceived a decrease, 19.4% perceived an unpredictable change, and 0.5% can't tell if there is a change in the local temperature during the night. Considering the local temperature during the day, 43.3% of the farmers perceived an increase, 32.7% perceived a decrease, 23.4% perceived an unpredictable, and 0.5% of them can't tell whether the local temperature has changed during the day. In contrast, the majority of AEWs perceived an increase in the local temperature during the day (60.0%) and night (63.3%). In confirmation, increasing temperatures since 1960 in Sierra Leone have been reported in the UNDP Climate Change Country Profile for Sierra Leone [8]. Figure 6 presents the results of the perception of farmers and AEWs on changes in the harmattan period during the last 30 years. The results show that 57.4% of the farmers observed a shortened, 31.6% observed a milder, 8.2%, observed no changes, 1.3% observed a prolonged, 1.1% observed a harsher, and 0.3% can't tell the changes in the harmattan period. Similarly, 53.3% of the AEWs observed a shortened harmattan period whilst 20.0% observed a prolonged one, 16.7% observed a milder and 10.0% can't tell whether the harmattan period has changed. Temperature trends from 1991 to 2020 in Moyamba district Linear relationship between year and temperature Figure 7 presents the results of the linear relationship between year and temperature distribution in the Moyamba district from 1991 to 2020. The results show a long-term fluctuation in the seasonal and annual distribution of mean temperature. The R-squared (R 2 ) statistic shows that the regression model accounted for 11.9%, 20.03%, and 16.79% of the variation in the dry season, rainy season, and annual mean temperature distribution respectively. The low R 2 values indicate a weak relationship between the mean temperature and year meaning the regression model accounts for less of the variance in the mean temperature and the data points of the mean temperature are widely dispersed around the regression line. The linear trend lines showed an upward trend in the rainy season, dry season, and annual mean temperature distribution. Temperature increase in Sierra Leone could be associated with the decreasing annual rainfall Atmospheric and Climate Sciences  phenomenon which is related to the variation in precipitation concentration intensity due to the interaction between the laden northeast and south-eastern winds [13]. High temperatures in recent times have also been reported to cause the insurgence of new insect pests and a high incidence of existing insect pests which are severely reducing crop yields to insect damage [10].
Significance and magnitude of the trend Table 4 presents the results of the Mann-Kendall test for the monthly, seasonal and annual time series for temperature in the Moyamba district from 1991 to 2020. At a 95% confidence limit, the results showed positive values of the MK-Stat and Z-stat for the monthly, seasonal, and annual temperature time series which indicates an upward trend in the mean temperature. However, the trend for the mean temperature was only significant (p < 0.05) for the months of May, July, August, September, October, and November, the rainy and dry season, and the annual time series, The result of an upward trend in the mean temperature is in agreement with the findings of the perception of farmers and Note: Negative and positive MK-Test, z-test and Sen's slope figures denote a decreasing and increasing trend respectively. A Z-stat figure of less than 1.96 and p-values greater than 0.05 indicate a non-significant trend at a 95% confidence limit.
AEWs on the changes in the local temperature in Moyamba district as a considerable number of the farmers and AEWs observed an increase in the local temperature during the night and day. The results of Sen's slope estimator showed that the trend of the rainy season, dry season, and annual mean temperature increased by 3.57%, 2.96%, and 3.18% respectively. The significant and upward trends of the seasonal and annual temperature in the country are a cause for concern as we will be expecting hotter years in the future which may be conducive to the insurgence of new pests and an increase in the incidence of pests that may considerably reduce crop yields [10]. Regarding the harmattan period, the trend of mean temperature during the harmattan months of November, December, January, and February also increased by 2.89%, 3.59%, 1.63%, and 1.60% respectively indicating a milder harmattan period. This result is in agreement with the perception of a substantial number of farmers and AEWs who observed a milder harmattan period in the district. Table 5 presents the results of the negative effects of extreme weather events on crop production in the Moyamba district. The results show that 84.8% of the farmers interviewed revealed that, adverse weather events have affected the major crops they cultivate. Among the crops mentioned by the farmers, rice (37.6%), followed by cassava (19.5%), and groundnut (24.2%) were the three main crops seriously affected by adverse weather events. Sweet potato (1.4%), maize (5.6%), yam (1.5%), pepper (7.4%), cowpea (1.4%), and pineapple (0.7%) were less affected by the negative effect of adverse weather conditions. Figure 8 presents the results of the perception of farmers on the adverse weather events that affect crop production in the Moyamba district. The results show that 48.1% of their crops were affected by drought, 24.5% by high temperature, 12.2% by irregular rainfall, 9.0% by flooding, 3.9% by erosion, 1.1% by storms/strong winds, and 1.1% by low temperatures. The results of previous studies have shown that increased temperature, and uneven and off-season rainfall leading to drought occurrences are the key negative impact of climate change on crop production [23].   Figure 9 presents the results of the estimated yield loss caused by extreme weather events in the Moyamba district. The magnitude of yield loss was estimated under three categories (above 60%, 40% -60%, and below 40%). The results show that the great majority (73.5%) of the interviewed farmers are estimated to have been experiencing more than 40% of yield loss due to extreme weather events. The high yield loss due to adverse weather events may have negative consequences on household food production in the district. Figure 10 presents the results of the coping strategies of farmers in response to adverse weather events. The results show that among the 339 farmers who experience adverse weather events, 41.9% of them do not take any action whilst 33.0% of the farmers change their crop production activities, 15.3% shift from on-farm to off-farm work, 3.5% use savings or borrow money, 3.5% reduce expenses at home and 2.7% sold household assets to get money in response to the negative effects of adverse weather events. Farmer's strategy in response to adverse weather events has been reported in previous studies [23]. Among the farmers who changed their crop production practices, 67.0% of them incurred an additional expenditure of SLL 879,640 (€73.32) ranging from SLL 250,000 (€20.84) to SLL 2,000,000 (€166.70).

Farmers' Knowledge and Use of Climate-Sensitive Agricultural
Practices Figure 11 presents the results of the farmer's knowledge and use of climatesensitive agricultural practices in the Moyamba district. The results show that a greater majority (77.5%) of the farmers interviewed know CSA practices whilst 22.5% do not know CSA practices. Among the farmers who know CSA practices, 34.8% of them are currently using CSA practices on their farms whilst 65.2% of them are not using CSA practices on their farms which can be attributed to their lack of technical and or financial capability to implement the CSA on their farms. Figure 12 presents the results of the type of CSA practices currently used by farmers in the Moyamba district. Among the farmers currently using CSA practices Atmospheric and Climate Sciences    on their farms, the majority of them (51.4%) are practicing intercropping. Results from previous studies have shown that farmers practice intercropping as a climate change adaptation strategy to increase production and land use efficiency [24] and crop diversification [25]. Besides intercropping, 44.8% of farmers are also currently planting short-duration crops, 44.0% of them use organic fertilizer, 42.8% of them practice minimum tillage, and 37.5% of them adapt their cropping calendar to prevailing weather conditions. Results from previous research studies have shown that farmers need to plant at different dates as a coping strategy to adapt to the erratic pattern of rainfall, and in cases where their crops repeatedly failed, they need to replant their fields many times depending on the commonly forecast weather indicators [26] and the use of weed residues to replenish soil fertility through minimum soil tillage practices and reduces surface runoff and erosion [27]. In addition, 32.2% of the farmers are using/ planting drought-resistant crops, 20.2% practicing in-field rainwater harvesting, 36.5% are practicing mulching, 16.2% are practicing reduced rates of the application of agrochemicals, 26.0% practice integrated pest management (IPM) and 31.2% practice integrated soil fertility management. Table 6 presents the results of the Kendall test of concordance in ranking the capacity-building needs of farmers and AEWs for climate change adaptation and mitigation in the Moyamba district. The results show that Kendall's concordance (Kendall's W) indicates a weak degree of agreement among the farmers (0.144) and AEWs of MAF Extension Workers (0.245) in the ranking of their capacity-building needs for climate change adaptation and mitigation at 1% probability level in the study area meaning the ranking of the capacity building needs among the farmers and the AEWs of MAF was however inconsistent. Access to climate-resilient crop varieties was ranked 1 st , training on CSA practices was ranked 2 nd and access to meteorological Information was ranked 3 rd by the farmers whilst training on CSA practices was ranked 1 st , access to climate-resilient crop varieties was ranked 2 nd , and access to meteorological information was ranked 3 rd by the AEWs of MAF. Previous studies have identified numerous ways in which governments, agricultural research centers, and other institutions could help smallholder farmers adapt to climate change, among these needs are the provision of climate-resilient crop varieties and training on CSA [28]. However, the climate change adaptation and mitigation needs of farmers differ when compared to the MAF Extension Workers. For instance, training on CSA was ranked 1 st by the MAF Extension Workers whilst climate-resilient crop varieties were ranked 1 st by the farmers. This result shows that farmers give more importance to inputs that are tangible for addressing the challenges posed by climate change whilst AEWs of MAF give more importance to training on CSA practices which are intangible. Therefore, project support to farmers and AEWs of MAF should be demand-driven and the strategies for addressing the challenges posed by climate change should be tailored according to their needs. Given the fact that both farmers and AEWs ranked 3 rd the access to meteorological information, means that this option/solution is given lower consideration in the ranking of their priority needs. This may be attributed to their lack of knowledge in understanding the implications of meteorological information on their agricultural production and livelihoods.

Conclusion
Both studies show evidence of climate change in the Moyamba district from 1991 to 2020 even though the observed changes in rainfall and temperature by the farmers and AEWs partially agree with the results of the analyzed meteorological data. Both studies show that the rainy season starts late and ends late and the annual rainfall amount is decreasing. On the contrary, the results of the analyzed meteorological data show that the length of the rainy season has not changed whilst the farmers and AEWs perceived a shorter rainy season period.
Both studies show that the local temperature has increased during the last 30 years. Changes in rainfall and temperature as perceived by farmers have resulted in adverse weather events. Droughts, high temperatures, and irregular rainfall as the three major extreme weather events that farmers perceived to have contributed to a substantial loss in the yields of the major crop crops cultivated on smallholder farms in the district. The increase in temperature contributes to the high incidence of pests and diseases. A substantial number of farmers take no action in response to the negative effects caused by climate change due to their lack of knowledge, technical or financial capacity to implement climate-smart crop production (CSA) practices. Even though few farmers are practicing some CSA practices (intercropping, short duration crops, organic fertilizer, minimum tillage, and adapting their cropping calendar) on their farms in response to the negative effects of extreme weather events, there is an urgent need to build the capacity of farmers and AEWs to improve crop production and adapt to and mitigate the negative impacts of climate change. The most priority need of farmers is to support them with climate-resilient crop varieties i.e. short duration and drought resistant varieties of the major crops (rice, cassava and groundnut) cultivated in the district, whilst the AEWs need training on CSA practices for the dissemination of appropriate extension messages for climate change adaptation and mitigation on smallholder farms.