Abstract

The livelihoods of the majority of people in semiarid areas of developing nations are based on rain-fed agriculture. In the wake of climate variability and change, communities in these regions are the most vulnerable because of their limited capacities to adapt to environmental changes. Smallholder farmers in the study area, Lower Gweru in central Zimbabwe, ascertain that they have observed changes in some rainfall and temperature patterns. These changes include higher temperatures, an increased number of seasons without enough rainfall, and an increased frequency of droughts and lengths of dry spells. The aim of this study was to find out whether farmers’ perceptions are supported by mean and extreme event trends in observed historical climate data. Gweru Thornhill meteorological data were analyzed for significant trends. The analysis showed that temperatures are increasing significantly, consistent with farmers’ observations that temperatures are getting hotter. This study revealed that farmer perceptions on rainfall were not consistent with historical climatic trends. Thus, farmers in the Lower Gweru area may not be a very reliable source of long-term rainfall trends.

1. Introduction

Agriculture is among the sectors that are negatively affected by climate variability and change (Boko et al. 2007). Climate variability and change are departures from the mean climate of a locality. The mean climate of a place is the unweighted temporal average over a long period of time (Arguez and Vose 2011). Future changes indicate an increase in the frequency and intensity of extreme climate events, such as heat waves, extreme cold spells, droughts, and floods (Niang et al. 2014). Such changes are likely to have bigger and negative impacts on agricultural productivity particularly in Africa, where communities rely heavily on rain-fed agriculture and have limited capacity to adapt (Boko et al. 2007; UNFCCC 2006). The understanding of how such climate extremes are changing is vital for planning appropriate adaptation measures (IPCC 2012; Aguilar et al. 2009). Reduced yields as a result of unfavorable and/or changing weather and climate leave farmers vulnerable to food insecurity because of communities’ limited capacities to adapt to environmental changes. During the last three decades or so, the southern Africa region has experienced frequent intense droughts (Vogel et al. 2010; Rouault and Richard 2005) that have had a negative impact on regional food security. A few studies, using daily data, have been carried out in Zimbabwe (e.g., Makuvaro 2014; Mazvimavi 2010; Aguilar et al. 2009; New et al. 2006) to determine the trends in precipitation and temperature extremes.

Zimbabwe is a land-locked country situated in southern Africa. The atmospheric circulation over southern Africa is dominated by the position and strength of the subtropical anticyclones of the Atlantic and Indian Oceans, and the position of the intertropical convergence zone (ITCZ). The rainfall mostly occurs as a result of the northward and southward migration of the ITCZ and the westerly cloud bands that result from the tropical temperate troughs (TTTs) (Mason and Jury 1997; Usman and Reason 2004; Hart et al. 2010). The moisture supply is mainly from the southwest Indian Ocean (Reason 2007). During drier seasons, anticyclonic circulation patterns dominate the regional airflow and the ITCZ remains anchored over the north. Southern Africa is also affected by tropical cyclones (TCs) that develop in the Indian Ocean. Tropical cyclones may bring in some moisture when they fall over land or dry the region when they remain in the Mozambique Channel (Matarira 1990). The rainfall pattern of southern Africa (including Zimbabwe) is affected by the ENSO signal (Rouault and Richard 2005; Reason and Jagadheesha 2005). When there is an El Niño, the country mostly experiences droughts, while a La Niña often results in favorable rainfall conditions.

The major challenge in trying to solicit farmers’ perceptions is to ascertain how long they recall yesteryear climatic conditions. Farmers whose perceptions were considered in this study were mostly in the 51–70-yr-old age group; hence, the researchers considered it reasonable to work with a climate record of 40 years. The long climate record is also ideal for picking up long-term trends [World Meteorological Organization (WMO) requires a period of at least 30 years]. According to Mertz et al. (2009), rural communities in eastern Saloum, Senegal, concurred that temperatures were increasing throughout the year with cold periods becoming shorter and hot ones becoming longer. Hachigonta et al. (2008) argue that information on specific aspects of a rainy season, such as its start and end and the nature of the wet and dry spells within it, is vital to farmers’ decision-making processes. Therefore, it is important to relate trends in the mean climate variables and extreme events with farmers’ perceptions on climate variability and change, as this will provide an appropriate basis for possible adaptation measures.

According to Mubaya (2010), about 90% of the farmers in Lower Gweru and Lupane, two semiarid districts of Zimbabwe, were fully aware of climatic variability and change in their surroundings. Case studies and interviews conducted in the Lower Gweru area in Gweru District revealed that farmers perceived that climate variability and change were taking a toll on their agricultural activities (A. Munodawafa et al. 2009, unpublished manuscript). The farmers specifically cited droughts, flooding, prolonged dry spells, and wet spells as negatively impacting their farming activities. Excessive rains resulted in increased livestock diseases, lower crop yields as a result of nutrient leaching, breaking down of dam walls, and increasing silt in rivers. On the other hand, they noted that droughts and prolonged dry spells resulted in the nonperennial flow of big rivers, reduced yields, and dried-up wells. Other studies have also established that rural farmers and communities in Africa are indeed aware of climate variability and change in their areas (Sanfo et al. 2014; Moyo et al. 2012; Mongi et al. 2010; Akponikpè et al. 2010; Gbetibouo 2009; Mertz et al. 2009; Maddison 2007; Nhemachena and Hassan 2007). Generally, in these studies farmers perceived a rise in temperature and drier conditions. In Nhemachena and Hassan (2007), the farmers from South Africa, Zambia, and Zimbabwe also concurred that there were significant changes in the start and end of seasons and changes in the frequency of droughts. Sanfo et al. 2014, Moyo et al. 2012, Mongi et al. 2010, Akponikpè et al. 2010, Gbetibouo 2009, and Maddison 2007 verified farmers’ perceptions on climate variability and change using historical data. In some of the countries where the investigations were carried out, the historical data were in agreement with farmers’ perceptions that both rainfall and temperatures had changed; however, in other case studies, there was no evidence that these weather variables had changed.

Smallholder farmers’ perceptions on climate variability and change are vital in determining adaptation strategies. Significant changes in climate, rainfall, and temperature in particular have a direct influence on agricultural productivity and therefore food and nutrition security. The objective of this study was to verify smallholder farmers’ perceptions on climate variability and change, using historical daily climate data. Similar work done in Zimbabwe focused on agroecological regions (AER) IV and V, which receive low rainfall amounts (less than 650 mm per annum; Vincent and Thomas 1962) and experience a high mean annual temperature of 20.5°–30°C (Muchadeyi et al. 2007). The agroecological regions in Zimbabwe are a classification of the agricultural potential—from AER I, which represents the highest altitude and wettest area, receiving more than 1000 mm of rainfall per year, to AER V, which receives the lowest rainfall, amounting to less than 450 mm yr−1 and is the driest and the hottest. This work looks at Lower Gweru, which is in AER III, which receives 650–800 mm per annum, has a relatively longer growing season (a mean median of 131 days), and a lower mean temperature than AERs IV and V (Vincent and Thomas 1962). The rainfall season in Zimbabwe normally starts in October and ends in March/April, and the peak rainfall period is December–February (DJF). The mean monthly rainfall variation for Gweru is shown in Fig. 1. In this study, daily data were used, since they are quicker at reacting to level shifts and changes in trends. While most of the research on the same subject of validating farmers’ perceptions of climate variability and change used mean climatic records, our study uses both mean and extreme climatic variables.

Fig. 1.

Mean monthly rainfall for the Gweru meteorological station.

Fig. 1.

Mean monthly rainfall for the Gweru meteorological station.

2. Materials and methods

a. Farmers’ location and their perceptions on climate variability and change

Perceptions of farmers from two wards (Mudubiwa and Nyama) in the Lower Gweru communal area on climate variability and change were compared with trends in average and extreme historical climatic data for the Gweru meteorological station. The communal area is situated about 45 km northwest of Gweru, in Gweru District. It lies in AER III, in semiarid central Zimbabwe. The two wards have an altitude of 1200–1345 MSL.

Farmer perceptions used in this study were obtained from a baseline survey on “Building Adaptive Capacity to Cope with Increasing Vulnerability due to Climate Change” conducted in the two wards during the 2008/09 season, under the International Development Research Centre Climate Change and Adaptation in Africa (IDRC CCAA) project (Mugabe et al. 2010). In that survey, Nyama and Mudubiwa wards (a ward is a geographical location with six–eight villages in it) were conveniently sampled for their proximity to the main road and therefore easy accessibility. While Nyama ward is predominated by wetlands and thus has a high water table, Mudubiwa is situated on high ground and therefore has a relatively low water table. Farmers in this area practice subsistence farming and market gardening.

Three villages were conveniently selected from each ward, and systematic random sampling was employed to come up with 30 households per village, bringing the total number of interviewed farmers to 180. Focus group discussions and questionnaires were the main methods of getting information from the farmers. In focus group discussions, farmers were put in groups of 8–15. Further grouping of farmers by age and gender was done so as to capture different responses from different farmers’ experiences on climate variability and change and to eliminate bias by dominant speakers from large groups. The three age groups used were 30–40 years, 41–50 years, and 51 years and older. Questions that farmers were asked included their knowledge of climate variability and change and their causes, signs they use to validate the changes, and start and cessation dates of rainfall, as well as adaptation strategies that they were undertaking.

The Lower Gweru smallholder farmer perceptions as recorded from the baseline survey were as follows: increased seasons without enough rainfall, increased floods, rains starting late and ending early, extremes in temperatures, long dry spells, and rains coming earlier. These perceptions are fairly similar to those identified for farmers in the semiarid Masvingo and Hwange Districts (Moyo et al. 2012) and Mangwe District (Tshuma and Mathuthu 2014) of Zimbabwe.

b. Meteorological data and analysis

Historical data for the Gweru meteorological station were obtained from the Department of Meteorological Services, Zimbabwe. Quality control and homogeneity tests were performed on the daily meteorological data.

The Statistical and Regional Dynamical Downscaling of Extremes for European Regions (STARDEX) software (http://www.cru.uea.ac.uk/projects/stardex) (Haylock et al. 2006) was used to determine trends in mean and extremes of the daily rainfall and temperature data for the Gweru meteorological station. Selected indices for average and extreme temperature and rainfall (Table 1) trends were compared to farmer perceptions. These indices were linked to extreme events, such as floods and droughts (wet and dry spells), and heat waves. In this study, a wet day was defined as a day on which 1 mm or more of rainfall was received and a dry day was considered as one that received less than 1 mm of rainfall. Usman and Reason (2004) also used this definition in their study for southern Africa dry spells.

Table 1.

Rainfall and temperature indices used in the study and their definitions.

Rainfall and temperature indices used in the study and their definitions.
Rainfall and temperature indices used in the study and their definitions.

For the start and cessation of the season, proxy indices were used as a result of the absence of ones that fully describe these parameters of the season. It seems there are no standard criteria for deciding the start and end of a season as revealed by the variations in definitions used by different practitioners (e.g., agronomists, agroclimatologists, and hydrologists) (Sanfo et al. 2014; Boyard-Micheau et al. 2013; Fosu-Mensah 2012; Hachigonta et al. 2008; Reason et al. 2005). A definition commonly used by agroclimatologists for start of a season is “the first wet day of a spell receiving a given rainfall amount and not followed by a long dry spell during the subsequent weeks” (Boyard-Micheau et al. 2013, p. 8916). Using these criteria, the definition for the start of a season adopted by Reason et al. (2005) in southern Africa is 25 mm in the first two pentads followed by 20 mm of rainfall in the next 20 days. Dimes et al. (2009) point out that a sowing criterion of 20 mm of rainfall received within 5 days during a particular sowing window is common. In Zimbabwe, the sowing window can be taken to be the period between 15 November and 15 January. Days with rainfall greater than 10 mm and the greatest amount of rainfall received during 5 consecutive days for September–November (SON) and March–May (MAM) were considered as the start and end of a season, respectively.

All computations for the mean and extreme indices were done relative to the base period 1962–82 for the Gweru climate data of 1962–2005.

c. Indices that were used to validate the farmer perceptions

Rainfall and temperature indices used in the study were measures of intensity, frequency, and proportion of the total (Table 1). Rainfall-related indices used were the mean climatological precipitation, the mean wet spell lengths, the number of days precipitation was greater than 10 mm, the greatest 10-day total rainfall, the number of the longest wet and dry spells, the number of rainfall events greater than the 90th percentile, the proportion of total rainfall received from rainfall events greater than the 90th percentile, and the amount of rainfall per rainy day. Temperature-related perceptions were compared to trends in mean maximum temperature, maximum temperature above the 90th percentile (hottest day temperature), percentage of days with a maximum temperature above the 90th percentile (frequency of hot days), mean minimum temperature, minimum temperature above the 90th percentile (hottest night temperature), and percentage of days with a minimum temperature above the 90th percentile (frequency of hot nights) (Table 1).

3. Results and discussion

a. Historical meteorological data

1) Rainfall

The mean climatological precipitation showed a significant (p = 0.0385) positive trend for MAM (end of rainfall season) (Fig. 2a) and nonsignificant (p > 0.05) trends for DJF (peak summer season), SON (start of rainfall season), and JJA. The annual trend for this index was also not significant (p > 0.05). The results for DJF, JJA, and SON are not shown.

Fig. 2.

Time series variations with linear trend lines for the rainfall extreme indices for the Gweru meteorological station: (a) MAM seasonal anomalies, (b) MAM seasonal number of days with rainfall greater than 10 mm, (c) MAM seasonal greatest 10-day rainfall total, (d) MAM seasonal simple daily intensity, and (e) annual simple daily intensity.

Fig. 2.

Time series variations with linear trend lines for the rainfall extreme indices for the Gweru meteorological station: (a) MAM seasonal anomalies, (b) MAM seasonal number of days with rainfall greater than 10 mm, (c) MAM seasonal greatest 10-day rainfall total, (d) MAM seasonal simple daily intensity, and (e) annual simple daily intensity.

The number of days with precipitation greater than 10 mm (frequency measure), the greatest 10-day total rainfall, and the amount of rain per rainy day (both intensity measures) revealed significant positive trends for MAM. The three indices showed increases of 0.0612 days yr−1, 0.9395 mm yr−1, and 0.1092 mm (rain day)−1 yr−1, respectively, with p values of 0.0153, 0.0255, and 0.0139, respectively (Figs. 2b–d). The number of days with precipitation greater than 10 mm and the greatest 10-day total rainfall indices registered nonsignificant trends (p > 0.05) for DJF, JJA, and SON, and for the annual season (results for these are not shown in the figures). The amount of rain per rain day index also recorded a significant (p = 0.0033) positive annual trend of 0.0575 mm (rain day)−1 yr−1 (Fig. 2e). The amount of rain per rain day index also registered nonsignificant (p > 0.05) trends for DJF, JJA, and SON.

Nonsignificant trends (p > 0.05) in the mean wet spell length, the longest dry spells, and the longest wet spells (frequency measures) were obtained for all seasons and for the year. Makuvaro (2014) also found nonsignificant trends in the mean wet spell length for the first half of the rainfall season, October–December (OND), and for the second half of the rainfall season, January–March (JFM), as well as for the annual trend for Bulawayo station in southwestern Zimbabwe. Aguilar et al. (2009) found a nonsignificant annual trend in the longest wet spell for some stations in Zimbabwe during the period 1955–2006. In a similar extreme rainfall study, Makuvaro (2014) established nonsignificant long wet spell trends for Bulawayo during OND and JFM for the period 1970–2007. In the same study, rainfall amounts and the frequency of heavy precipitation events showed no significant trends (p > 0.05). Mazvimavi (2010) also established no significant trends in heavy precipitation for several stations in Zimbabwe for intervals starting during the 1892–1941 period and ending in 2000. Aguilar et al. (2009) also noted a nonsignificant trend in consecutive dry days (longest dry spell) for Zimbabwe for the period 1955–2006, while Makuvaro (2014) obtained a significant increase in the longest dry spell for Bulawayo station, in western Zimbabwe, during OND.

The greatest 3- and 5-day total rainfall indices recorded significant positive trends (p < 0.05)—0.6493 mm per 3 days (p = 0.01) and 1.0171 mm per 5 days (p = 0.0187), respectively,—for MAM. These rainfall intensity measures registered nonsignificant trends (p > 0.05) for DJF, JJA, and SON, and for the year (results for these are not shown in the figures). The percentage of total rainfall from events greater than the long-term 90th percentile (proportion measure) and the number of events greater than the long-term 90th percentile (frequency measure) indices revealed nonsignificant trends (p > 0.05) for the seasons and for the year (the results for these are not shown in the figures).

Variations in nature of trends among the cited results (based on climate records) are mainly due to differences in seasons and the period of study considered, as well as the geographical coverage (single station vs average over several stations). Also, rainfall in Zimbabwe is highly variable both spatially and temporally, hence different outcomes in trend analyses are expected for different locations.

2) Temperature

The long-term time series for Gweru mean maximum temperature is indicative of a significant (p = 0.003) annual rise of 0.0212°C yr−1 for the 43-yr period of analysis (Fig. 3). Seasonal trends for this index were also significantly positive for JJA (p = 0.003) and SON (p = 0.0473), being 0.0245° and 0.0185°C yr−1, respectively (Fig. 3).

Fig. 3.

Time series variations with trend lines for the mean maximum temperature for the Gweru meteorological station (JJA, SON, and annual).

Fig. 3.

Time series variations with trend lines for the mean maximum temperature for the Gweru meteorological station (JJA, SON, and annual).

The maximum temperature above the 90th percentile (hottest day temperature) showed a significant (p = 0.0012) annual increase of about 0.0293°C yr−1 for the study period (Fig. 4a). All 3-month periods except MAM also showed significant positive trends: 0.0293°C yr−1 (p = 0.0453) for DJF, 0.0461°C yr−1 (p = 0.0005) for JJA, and 0.0247°C yr−1 (p = 0.0162) for SON (Figs. 4b–d, respectively). The percentage of days with maximum temperature above the 90th percentile (frequency of hot days) showed a significant (p = 0.0009) annual increase of 0.0021 days yr−1 (Fig. 5a), while for JJA and SON significant trends of 0.0026 (at p = 0.0002) and 0.0025 (at p = 0.0136), respectively, were obtained (Figs. 5b,c, respectively).

Fig. 4.

Time series variations with linear trend lines for the extreme (90th percentile) maximum temperature for the Gweru meteorological station: (a) annual, (b) DJF, (c) JJA, and (d) SON.

Fig. 4.

Time series variations with linear trend lines for the extreme (90th percentile) maximum temperature for the Gweru meteorological station: (a) annual, (b) DJF, (c) JJA, and (d) SON.

Fig. 5.

Time series variations with linear trend lines for the frequency of extreme (90th percentile) maximum temperature for the Gweru meteorological station: (a) annual, (b) JJA, and (c) SON.

Fig. 5.

Time series variations with linear trend lines for the frequency of extreme (90th percentile) maximum temperature for the Gweru meteorological station: (a) annual, (b) JJA, and (c) SON.

Unganai (1996) showed a similar (positive) annual trend of 0.6°C in mean maximum temperature and seasonal trends ranging from +0.1 to +0.8 for the Harare meteorological station, north of Zimbabwe, for the period 1897–1993. Makuvaro (2014) also found significant increases in maximum temperature above the 90th percentile (hottest day temperature) and in hot day frequency during winter (JJA), spring (SON), and for the year for Bulawayo station (southwestern Zimbabwe) for the period 1978–2007.

The mean minimum temperature showed a significant (p = 0.0002) positive annual trend of 0.0204°C yr−1 over the analysis period (Fig. 6a). All 3-month periods except DJF had positive trends: 0.0206°C yr−1 (p = 0.0365) for MAM, 0.0191°C yr−1 (p = 0.0310) for SON, (Fig. 6b) and 0.0286°C yr−1 (p = 0.001) for JJA (Fig. 6a).

Fig. 6.

Time series variations with linear trend lines for the mean minimum temperature for the Gweru meteorological station: (a) annual and JJA, and (b) MAM and SON.

Fig. 6.

Time series variations with linear trend lines for the mean minimum temperature for the Gweru meteorological station: (a) annual and JJA, and (b) MAM and SON.

The annual and seasonal (except SON) trends for minimum temperature above the 90th percentile (hottest night temperature) were significantly positive (Fig. 7). The annual change was 0.0187°C yr−1 (p = 0.0058) for the analysis period, while seasonal increases of 0.0452°C yr−1 (p = 0.0002), 0.021°C yr−1 (p = 0.0403), and 0.0157°C yr−1 (p = 0.0374) were noted for JJA, MAM, and DJF, respectively (Fig. 6), for the same period. The percentage of days with minimum temperature above the 90th percentile (frequency of hot nights) showed a significant (p = 0.0012) annual increase of 0.0018 and an increase of 0.0026 (p = 0.001) for JJA (Fig. 8). Similar to findings from this study, Unganai (1996) established positive annual and seasonal trends in mean minimum temperature for Harare station, while Makuvaro (2014) found significant increases in minimum temperatures above the 90th percentile (hottest nighttime temperatures) and in hot night frequency for SON and JJA, respectively, for Bulawayo station.

Fig. 7.

Time series variations with linear trend lines for the extreme (90th percentile) minimum temperature for the Gweru meteorological station: (a) DJF, (b) MAM, (c) JJA, and (d) annual.

Fig. 7.

Time series variations with linear trend lines for the extreme (90th percentile) minimum temperature for the Gweru meteorological station: (a) DJF, (b) MAM, (c) JJA, and (d) annual.

Fig. 8.

Time series variations with linear trend lines for the frequency of extreme (90th percentile) minimum temperature for the Gweru meteorological station: (a) annual, (b) JJA, (c) SON, (d) DJF.

Fig. 8.

Time series variations with linear trend lines for the frequency of extreme (90th percentile) minimum temperature for the Gweru meteorological station: (a) annual, (b) JJA, (c) SON, (d) DJF.

The number of days above the 90th percentile of a heat wave (the duration of the longest heat wave) showed a significant (p = 0.0018) increase of 0.0997 days yr−1, and of 0.053 days yr−1 for SON and 0.0253 days yr−1 for DJF (Fig. 9). The trends for the other seasons were insignificant (p > 0.05).

Fig. 9.

Time series variations with linear lines for the longest heat wave duration for the Gweru meteorological station: (a) annual, (b) SON, and (c) DJF.

Fig. 9.

Time series variations with linear lines for the longest heat wave duration for the Gweru meteorological station: (a) annual, (b) SON, and (c) DJF.

b. Climate variability

Rainfall variability

The rainfall anomalies reveal high variability as depicted in Fig. 2a for MAM. Similar results are obtained for the other seasons and for the annual series (not shown). All the precipitation and temperature indices considered in this study (Table 1) also reveal high temporal variations (Figs. 2, 4, 5, 7, 8, 9). Figures 3, 6 show the temporal variations for the mean maximum and minimum temperatures, respectively.

c. Comparison between farmer perceptions and historical climate data

1) Increased seasons without enough rainfall

To verify the farmers’ perception that there were increased seasons without enough rainfall, the mean climatological precipitation, the longest wet spell, the mean wet spell lengths, the number of days of precipitation with greater than 10 mm, the greatest 10-day total rainfall, and the amount of rain per rainy were considered. In summary, all the mean and extreme indices used to verify the farmers’ perception that there were increased seasons without enough rainfall showed nonsignificant (p > 0.05) trends except for the positive trends during MAM. These positive trends for the end of season are indicative of rainfall increases contrary to farmers’ perceptions that there are increased seasons without enough rainfall. Increased rainfall amounts at the end of the growing season may lead to crop loss due to infestation of crop products by bacterial and fungal diseases while the crop is still in the fields awaiting harvesting. On a positive note, more rainfall toward the end of the season will provide conducive conditions for autumn plowing in preparation for the establishment of the subsequent crop. Contrary to results obtained for Lower Gweru in this study, Jiri et al. (2015) found that farmer perceptions of reduced annual rainfall agreed with observed climate records for Chiredzi District in southeastern Zimbabwe for the period 1980–2011.

Trends in the mean and extreme rainfall indices, considered above, did not support Lower Gweru farmers’ perception that the number of seasons without enough rainfall had increased.

2) Increased floods

The farmers’ perception of increased frequency of floods was validated using the number of the longest wet spells, the number of rainfall events greater than the 90th percentile, the proportion of total rainfall received from rainfall events greater than the 90th percentile, and the amount of rainfall per rainy day. The lack of significant (p > 0.05) annual and seasonal trends in the number of the longest wet spells, the number of rainfall events greater than the 90th percentile, and the proportion of total rainfall received from rainfall events greater than the 90th percentile does not support farmers’ perception that floods had increased. The change in amount of rainfall per rainy day was also not significant except for MAMm which showed a positive trend of 0.1092 mm (rain day)−1 yr−1 (p = 0.0139). Generally, the results lacked evidence of an increase in heavy precipitation except toward the end of the rain season (MAM period). An increase in floods suggests a need to introduce more efficient flood warning systems to alert the farmers of any flooding should it occur. Flooding would also impact negatively on the food and nutrition security of the farmers, since it may result in poor yields because of leaching and other plant growth problems associated with waterlogging.

3) Rains start late and end early

Proxy indices for the start and end of a season were used, as there were no climatological indices directly linked to definitions of “start” and “end” of the growing season (see section 2b). The days with rainfall greater than 10 mm, the amount of rain per rain day, and the greatest amount of rainfall received during 10 consecutive days for SON—the most likely rainfall onset period—showed nonsignificant (p > 0.05) trends. Thus, there is insufficient evidence of a delayed start of the season for the period of analysis (1962–2005), as revealed by nonsignificant (p > 0.05) trends in these proportion, frequency, and intensity of rainfall-measuring indices. Simelton et al. (2013) also found no evidence to support perceptions of a late start to the season by farmers in Botswana and southern Malawi. For cropping purposes a significant decrease in rainfall during the likely period for the start of the rains is an indicator of a delayed onset of the season, and farmers start planting later only when sufficient rains are received. The farmers also perceived that the rains were ending early. To verify this, the study considered the days with rainfall greater than 10 mm, the amount of rain per rain day, and the greatest amount of rainfall received during 10 consecutive days for the MAM period of the rainfall season. All the three indices (measures of proportion, frequency, and intensity) showed significant (p = 0.0153, 0.0139 and 0.0255, respectively) positive trends (0.0612 days yr−1, 0.1092 mm (rain day)−1 yr−1, and 0.9395 mm day−1 yr−1, respectively) (Figs. 2b–d). Thus, these indices contradict farmers’ perception that seasons were ending early. If rains start late and end early, as perceived by most of the farmers in Lower Gweru, it means shortening of the growing season. However, the analysis from this study indicates no change to the start of the season and a delayed end of the season, which means an extended growing season. A combination of no change in rainfall frequency and intensity during the greater part of the growing season and an extended growing season (established in this study) will probably leave farmers to continue growing the same crops but also to grow medium-maturity varieties as opposed to be confined to short-season varieties. The extended season could also allow for double cropping by establishing a second early maturing crop, like sugar beans or cowpeas.

4) Long dry spells

The seasonal and annual trends in the consecutive dry days (frequency measure) were analyzed. The longest dry spells trends were not significant (p > 0.05), contrary to what farmers in Lower Gweru perceived. Thus, the results of this study show no evidence of longer dry spells than the long-term average for Lower Gweru. Increased frequency of long dry spells will compel farmers to embark on strategies for managing dry spells, such as growing drought-tolerant crops, employing moisture conservation techniques, and putting more emphasis on livestock enterprises.

Overall, there was limited agreement between farmers’ perceptions and historical mean and extreme climate data with respect to rainfall amount and patterns (Table 2). The lack of match between the two sources of information found in this study is probably due to the differences in the reference periods. Whereas extreme climate trends used in this study are for a relatively long time frame (1962–2005), the farmers may not have considered a long period such as that used by the researchers in trying to remember past years. Their responses could have been based on relatively recent years and/or on years that were extremely “bad” or “good”; furthermore, as alluded to by previous researchers (e.g., Sanfo et al. 2014; Simelton et al. 2013, Gbetibouo 2009), these bad and good years are not named as such on the basis of climatic factors but rather on agricultural output, which is affected by a plethora of factors. During the period of the study, southern Africa has been affected by bad years (1982/83, 1991/92, 2002/03, 2004/05, 2005/06, and 2007/08 droughts), as well as good years (1979/80, 1997/98, and 1999/2000 wet rainfall seasons). It appears the farmers are more concerned or conscious about inter- to intraseasonal rainfall variability than climate change. This assertion is confirmed by the results of analyses by some researchers (e.g., Sanfo et al. 2014; Simelton et al. 2013; Moyo et al. 2012), which showed that farmers’ perceptions were true for particular years, especially those that received below- or above-average rainfall.

Table 2.

Summary of comparisons between farmer perceptions and historical climate data.

Summary of comparisons between farmer perceptions and historical climate data.
Summary of comparisons between farmer perceptions and historical climate data.

5) Extreme temperatures

The mean and extreme temperature indices computed in section 3a(2) show that temperatures have been increasing as noted by the farmers. The warming of temperatures also agrees with Aguilar et al. (2009) and New et al.’s (2006) findings of small increases in maximum temperature trends for Zimbabwe for the periods 1955–2006 and 1961–2000, respectively. Other studies in Zimbabwe also showed upward trends in maximum temperature (e.g., Jiri et al. 2015 for Chiredzi; Moyo et al. 2012 for Hwange and Masvingo). Elsewhere in Africa, similar results were reported (e.g., Sanfo et al. 2014 for Burkina Faso; Mongi et al. 2010 for Tanzania; Gbetibouo et al. 2009 for West Africa). From an agricultural viewpoint, the increase in temperature may directly influence the biophysical processes of plants and animals, leading to poor productivity. The increase in the extreme temperature may increase pest and disease incidences and may increase rates of evaporation, leading to increased crop water requirements and reduced water availability. Thus, these changes could contribute to reduced agricultural productivity and threaten food security in the area. Temperature increases, as noted by farmers and also verified in this study, call for development of crop varieties and animal breeds that are heat stress tolerant.

There is much agreement between farmers’ perception and historical climate data regarding an increase in temperature. Other researchers also found similar results (e.g., Jiri et al. 2015; Sanfo et al. 2014; Simelton et al. 2013; Mongi et al. 2010; Moyo et al. 2012; Maddison 2007). The reduced coherence between the different sources of climate information (farmer perceptions and climate records) pertaining to rainfall, compared to temperature, could be due to less variability in temperature compared to rainfall.

4. Conclusions

This study sought to compare the Lower Gweru farmer perceptions on climate variability and change as of 2008 with trends in historical climate records (1962–2005). The mean and extremes from the historical meteorological data for both rainfall and temperatures are characterized by high interannual variability. Some of the trends in these mean and extreme values were significant (p < 0.05), while others were nonsignificant.

Most of the rainfall frequency measures (the mean wet spell length, the longest dry and wet spells, the number of events greater than the long-term 90th percentile) used in this study registered nonsignificant seasonal and annual trends. However, precipitation greater than 10 mm revealed significant (p < 0.05) positive trends for the end of season. Intensity measures of rainfall (the greatest 10-day total rainfall, the amount of rain per rain day, and simple daily intensity) were mostly nonsignificant except for the end of rainfall season window. The proportion measure for rainfall used—that is, the percentage of total rainfall from events greater than the long-term 90th percentile—was nonsignificant. The local farmers’ perceptions on rainfall were that they had experienced increased seasons without enough rainfall, increased floods, rains starting late and ending early, and long dry spells. The major conclusion from this study was that the Lower Gweru farmers’ perceptions on rainfall amount and pattern did not correspond with climatic trends. Thus, farmers in Lower Gweru may not be a reliable source of long-term changes in rainfall, but they could provide reliable information on inter- and intraseasonal rainfall changes. It appears that for a variable climate element like rainfall, studies on comparisons of perceptions and trends need to be done on smaller time scales, for example, decadal periods rather than on long periods. This is particularly important when such studies seek communities’ responses.

On the other hand, their perceptions on temperature changes were in total agreement with the meteorological trends. The mean maximum temperature and the frequency of hot days were characterized by significant (p < 0.05) positive trends for the year and for the dry seasons, though the wet seasons showed nonsignificant trends. The hottest day temperature showed warming for most of the seasons and for the year. The longest heat wave duration revealed significant (p < 0.05) increases in the second half of the year and on an annual basis. Gweru station meteorological data showed significant (p < 0.05) increases in the mean minimum temperature over the years and for most of the seasons. The hottest nights and their frequency registered significant (p < 0.05) increases annually and for most of the seasons. The Lower Gweru farmers had noted extreme temperatures. Meteorological data revealed that both the mean and extremes of maximum and minimum temperatures had generally showed signs of warming.

The study recommends that a broader study on verifying farmer perceptions using historical climatic data across the country be carried out. Findings from such studies would help smallholder farmers, policy makers, and researchers in finding ways to combat climate variability and change.

Acknowledgments

The authors thank staff members from the Department of Meteorological Services, Zimbabwe, for providing the meteorological data. We are also grateful to Dr. Steven Crimp of the Commonwealth Scientific and Industrial Research Organisation (CSIRO), Canberra, Australia, who facilitated access to the STARDEX software used in this study and for his assistance with the interpretation of the various extreme climate indices. The farmers’ perceptions are an output from a project funded by the International Development Research Centre Climate Change and Adaptation in Africa (IDRC CCAA) and DFiD project (Grant 104144). The views expressed are not necessarily those of CCAA/IDRC/DFiD.

REFERENCES

REFERENCES
Aguilar
,
E.
, and Coauthors
,
2009
:
Changes in temperature and precipitation extremes in western central Africa, Guinea Conakry, and Zimbabwe, 1955–2006
.
J. Geophys. Res.
,
114
,
1955
2006
, https://doi.org/10.1029/2008JD011010.
Akponikpè
,
P. B. I.
,
P.
Johnston
, and
E. K.
Agbossou
,
2010
: Farmers’ perception of climate change and adaptation strategies in Sub-Saharan West-Africa. Second Int. Conf. on Climate, Sustainability and Development in Semi-Arid Regions (ICID 2010), Fortaleza, Ceará, Brazil, IISD, ICID+18, 14 pp.
Arguez
,
A.
, and
R. S.
Vose
,
2011
: The definition of the standard WMO climate normal: The key to deriving alternative climate normals. Bull. Amer. Meteor. Soc., 92, 699–704, https://doi.org/10.1175/2010bams2955.1.
Boko
,
M.
, and Coauthors
,
2007
: Africa. Climate Change 2007: Impacts, Adaptation and Vulnerability, M. L. Parry et al., Eds., Cambridge University Press, 433–467.
Boyard-Micheau
,
J.
,
P.
Camberlin
, and
N.
Philippan
,
2013
:
Regional-scale rainy season onset detection: A new approach based on multivariate analysis
.
J. Climate
,
26
,
8916
8928
, https://doi.org/10.1175/JCLI-D-12-00730.1.
Dimes
,
J.
,
P.
Cooper
, and
K. P. C.
Rao
,
2009
: Climate change impact on crop productivity in the semi-arid tropics of Zimbabwe in the 21st century. Increasing the productivity and sustainability of rainfed cropping systems of poor smallholder farmers: Proceedings of the CGIAR Challenge Program on Water and Food International Workshop on Rainfed Cropping Systems, E. Humphreys and R. S. Bayot, Eds., CGIAR Challenge Program on Water and Food, 189–198.
Fosu-Mensah
,
B. Y.
,
2012
: Modelling maize (Zea mays L.) productivity and impact of climate change on yield and nutrient utilization in sub-humid Ghana. Ph.D. thesis, Kwame Nkrumah University of Science and Technology, ZEF Ecology and Development Series 87, 171 pp., http://www.zef.de/fileadmin/webfiles/downloads/zefc_ecology_development/eds_87_fosu-mensah_text.pdf.
Gbetibouo
,
G. A.
,
2009
: Understanding farmers’ perceptions and adaptations to climate change and variability: The case of the Limpopo Basin, South Africa. International Food Policy Research Institute Discussion Paper 00849, 36 pp., http://www.ifpri.org/publication/understanding-farmers-perceptions-and-adaptations-climate-change-and-variability.
Hachigonta
,
S.
,
C. J. C.
Reason
, and
M.
Tadross
,
2008
:
An analysis of onset date and rainy season duration over Zambia
.
Theor. Appl. Climatol.
,
91
,
229
243
, https://doi.org/10.1007/s00704-007-0306-4.
Hart
,
N. C. G.
,
C. J. C.
Reason
, and
N.
Fauchereau
,
2010
:
Tropical–extratropical interactions over southern Africa: Three cases of heavy summer season rainfall
.
Mon. Wea. Rev.
,
138
,
2608
2623
, https://doi.org/10.1175/2010MWR3070.1.
Haylock
,
M. R.
,
G. C.
Cawley
,
C.
Harpham
,
R. L.
Wilby
, and
C. M.
Goodess
,
2006
:
Downscaling heavy precipitation over the United Kingdom: A comparison of dynamical and statistical methods and their future scenarios
.
Int. J. Climatol.
,
26
,
1397
1415
, doi:.
IPCC
,
2012
: Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. C. B. Field et al., Eds., Cambridge University Press, 582 pp.
Jiri
,
O.
,
P.
Mafongoya
, and
P.
Chivenge
,
2015
:
Smallholder farmer perceptions on climate change and variability: A predisposition for their subsequent adaptation strategies
.
J. Earth Sci. Climatic Change
,
6
,
277
, https://doi.org/10.4172/2157-7617.1000277.
Maddison
,
D.
,
2007
: The perception of and adaptation to climate change in Africa. World Bank Policy Research Working Paper WPS4308, 51 pp., https://openknowledge.worldbank.org/handle/10986/7507.
Makuvaro
,
V.
,
2014
: Impact of climate change on smallholder productivity in Zimbabwe, using a modelling approach. Ph.D. thesis, University of the Free State, 296 pp.
Mason
,
S. J.
, and
M. R.
Jury
,
1997
:
Climatic variability and change over southern Africa: A reflection on underlying processes
.
Prog. Phys. Geogr.
,
21
,
23
50
, https://doi.org/10.1177/030913339702100103.
Matarira
,
C. H.
,
1990
:
Drought over Zimbabwe in a regional and global context
.
Int. J. Climatol.
,
10
,
609
625
, https://doi.org/10.1002/joc.3370100605.
Mazvimavi
,
D.
,
2010
:
Investigating changes over time of annual rainfall in Zimbabwe Afr
.
Hydrol. Earth Syst. Sci.
,
14
,
2671
2679
, https://doi.org/10.5194/hess-14-2671-2010.
Mertz
,
O.
,
C.
Mbow
,
A.
Reenberg
, and
A.
Diouf
,
2009
:
Farmers’ perceptions of climate change and agricultural adaptation strategies in rural Sahel
.
Environ. Manage.
,
43
,
804
816
, https://doi.org/10.1007/s00267-008-9197-0.
Mongi
,
H.
,
A. E.
Majule
, and
J. G.
Lyimo
,
2010
:
Vulnerability and adaptation of rain fed agriculture to climate change and variability in semi-arid Tanzania
.
Afr. J. Environ. Sci. Technol.
,
4
,
371
381
, https://doi.org/10.5897/ajest09.207.
Moyo
,
M.
,
B. M.
Mvumi
,
M.
Kunzekweguta
,
K.
Mazvimavi
,
P.
Craufurd
, and
P.
Dorward
,
2012
:
Farmer perceptions on climate change and variability in semi-arid Zimbabwe in relation to climatology evidence
.
Afr. Crop Sci. J.
,
20
(
Suppl. 2
),
317
335
.
Mubaya
,
C. P.
,
2010
: Farmer strategies towards climate variability and change in Zimbabwe and Zambia. Ph.D. thesis, University of the Free State, 259 pp.
Muchadeyi
,
F. C.
,
C. B. A.
Wollny
,
H.
Eding
,
S.
Weigend
,
S. M.
Makuza
, and
H.
Simianer
,
2007
:
Variation in village chicken production systems among agro-ecological zones of Zimbabwe
.
Trop. Anim. Health Prod.
,
39
,
453
461
, https://doi.org/10.1007/s11250-007-9050-0.
Mugabe
,
F. T.
, and Coauthors
,
2010
: Building adaptive capacity to cope with increasing vulnerability due to climate change. IDRC Final Tech. Rep. 104144, 175 pp.
New
,
M.
, and Coauthors
,
2006
:
Evidence of trends in daily climate extremes over southern and west Africa
.
J. Geophys. Res.
,
111
,
D14102
, https://doi.org/10.1029/2005JD006289.
Nhemachena
,
C.
, and
R.
Hassan
,
2007
: Micro-level analysis of farmers’ adaptation to climate change in southern Africa. International Food Policy Research Institute Discussion Paper 00714, 31 pp.
Niang
,
I.
,
O. C.
Ruppel
,
M. A.
Abdrabo
,
A.
Essel
,
C.
Lennard
,
J.
Padgham
, and
P.
Urquhart
,
2014
: Africa. Climate Change 2014: Impacts, Adaptation, and Vulnerability, V. R. Barros et al., Eds., Cambridge University Press, 1199–1265.
Reason
,
C. J. C.
,
2007
:
Tropical cyclone Dera, the unusual 2000/01 tropical cyclone season in the South West Indian Ocean and associated rainfall anomalies over southern Africa
.
Meteor. Atmos. Phys.
,
97
,
181
188
, https://doi.org/10.1007/s00703-006-0251-2.
Reason
,
C. J. C.
, and
D.
Jagadheesha
,
2005
:
A model investigation of recent ENSO impacts over southern Africa
.
Meteor. Atmos. Phys.
,
89
,
181
205
, https://doi.org/10.1007/s00703-005-0128-9.
Reason
,
C. J. C.
,
S.
Hachigonta
, and
R. F.
Phaladi
,
2005
:
Interannual variability in rainy season characteristics over the Limpopo region of southern Africa
.
Int. J. Climatol.
,
25
,
1835
1853
, https://doi.org/10.1002/joc.1228.
Rouault
,
M.
, and
Y.
Richard
,
2005
:
Intensity and spatial extent of droughts in southern Africa
.
Geophys. Res. Lett.
,
32
,
L15702
, doi:.
Sanfo
,
S.
,
J. P. A.
Lamers
,
M.
Mueller
, and
W.
Fonta
,
2014
: Farmers’ perception of climate change and variability versus climatic evidence in Burkina Faso, West Africa. Climate Change in Africa: Negotiations, Translations, and Socio-Political Implications, Bonn, Germany, Center for Development Research (ZEF), 16 pp., https://doi.org/10.13140/2.1.4614.7528.
Simelton
,
E.
, and Coauthors
,
2013
:
Is rainfall really changing? Farmers’ perceptions, meteorological data, and policy implications
.
Climate Dev.
,
5
,
123
138
, https://doi.org/10.1080/17565529.2012.751893.
Tshuma
,
N.
, and
T.
Mathuthu
,
2014
:
Climate change in Zimbabwe: Perceptions of smallholder farmers in Mangwe District
.
Int. J. Humanit. Soc. Stud.
,
2
(
5
),
318
325
.
UNFCCC
,
2006
: Climate change: Impacts, vulnerabilities and adaptation in developing countries. United Nations Framework Convention on Climate Change, 64 pp., https://unfccc.int/resource/docs/publications/impacts.pdf.
Unganai
,
L. S.
,
1996
:
Historic and future climatic change in Zimbabwe
.
Climate Res.
,
6
,
137
145
, https://doi.org/10.3354/cr006137.
Usman
,
M. T.
, and
C. J. C.
Reason
,
2004
:
Dry spell frequencies and their variability over southern Africa
.
Climate Res.
,
26
,
199
211
, https://doi.org/10.3354/cr026199.
Vincent
,
V.
, and
R. G.
Thomas
,
1962
: An Agricultural Survey of Southern Rhodesia. Part I: Agro-Ecological Survey. Federation of Rhodesia and Nyasaland, 217 pp.
Vogel
,
C.
,
I.
Koch
, and
K. V.
Zyl
,
2010
:
“A persistent truth”—Reflections on drought risk management in southern Africa
.
Wea. Climate Soc.
,
2
,
9
22
, https://doi.org/10.1175/2009WCAS1017.1.

Footnotes

© 2018 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).