Extreme weather events and crop diversification: climate change adaptation in Brazil

Climate variability and the frequency and intensity of extreme events are expected to increase in some areas of the world (IPCC 2023). In Brazil, the frequency and intensity of drought have increased in the northern and northeast regions (Avila-Diaz et al. 2023). According to Ballarin et al. (2023), a reduction in the total amount of rain and the number of humid days is expected, and an increase in the number of consecutive dry days in different climate scenarios is proposed by the IPCC Sixth Assessment Report. In addition, an increase in heavy rain is expected in the Brazilian South and Southeast regions (Da Silva Tavares et al. 2023). There is also evidence of the increased frequency of heatwaves around Brazil, especially in the Semiarid region and the Amazon (Sanches et al. 2023; Avila-Diaz et al. 2023; Ballarin et al. 2023), as well as more severe and intense cold waves in most of the southern region (Marengo et al. 2023).

Furthermore, climate variability considerably impacts agriculture, especially in countries that depend on agricultural production. There is evidence that abiotic stress due to adverse climatic conditions reduces the productivity of crops in the main agricultural products in the world (Mittler 2006; Boyer 1982). According to Ray et al. (2015), the global variability in the productivity of important crops, such as corn, rice, wheat, and soy, depends, around 32% to 39%, on climate variation. Thus, studies that analyze how extreme weather affects agricultural production should be highlighted because this can impact the agricultural economies. A negative relationship exists between temperature and the Gross Domestic Product (GDP) levels in tropical countries, which depend economically on agriculture (Newell et al. 2021). Regarding Brazil, summer droughts have caused a decline in the GDP growth rate of the Northeast region, and summer floods have negatively affected the Northeast, Central-west, and Southeast, which are essential agricultural sectors (Tebaldi & Beaudin 2016). The economy could be vulnerable to climate variability in a country where the agribusiness sector had an average share of approximately 24% of the Brazilian GDP in 2023 (CEPEA 2023).

However, the negative impacts on agriculture due to extreme weather can be reduced if actions to adapt to climate change occur. In this way, crop diversification helps to reduce observed or projected climatic risks (Heal et al. 2014). Therefore, diversification will be considered an adaptive strategy that helps reduce negative impacts and takes advantage of positive effects to maximize the farmer's well-being. Crop diversification promotes economic benefits, such as reducing the variability of farmers' income (Chavas 2023) and reducing rural poverty (Renard and Tilman 2019). Additionally, diversification promotes agroecological services because the increased richness of crop species distributed in space and time decreases the susceptibility to diseases, pests, and weeds, reduces soil erosion, and improves soil fertility (Huss et al. 2022). These positive effects are related to different types of crop diversification, such as intercropping, double-cropping, crop rotation, crop-livestock integration or mixed farming systems, crop-livestock-forest integration or agroforestry (Piedra-Bonilla et al. 2020a, b). These are considered technologies that sustainably enhance resilience and reduce climate risks. For example, crop rotation can decrease crop yield losses caused by drought (Renwick et al. 2021) and yield gains in cereals with high temperatures and low precipitation levels (Marini et al. 2020). Similarly, agroforestry helps increase soil water storage even in drought conditions (Zhao et al. 2022) and mitigates floods (Baldwin et al. 2022). Moreover, intercropping cereals with legumes can enhance crop yield and resilience in situations such as deficit irrigation (Aslam et al. 2020; Amanullah et al. 2021). Simulation studies show that crop-livestock systems would present greater resilience under future climate change conditions (Peterson et al. 2020; Delandmeter et al. 2024).

Regarding climate variables, most research indicates statistically significant effects on crop concentration. Research converges on the positive impacts of temperature variables on diversification, including their different units of measurement. However, precipitation shocks show ambiguous results (Seo and Mendelsohn 2008; Dillon et al. 2015; Asfaw et al. 2018; Arslan et al. 2018). These results depend on the geographical conditions of the regions. However, scant empirical evidence exists on how farmers respond to extreme weather, especially using the climate extreme indices. Thus, the present work intends to focus on the effect of extreme weather in the recent past and the future climate change scenarios on crop diversification.

Dealing with this topic in Brazil is especially important since the country is one of the world's main food producers and exporters (FAO 2018). Furthermore, Brazilian agriculture is heterogeneous around its regions. According to Piedra-Bonilla et al. (2020a, b), small farms (0 to 10 ha) in regions such as the South, Mid-West, and North are conducive to crop diversification. However, the Mid-West region, the area of grain production, tends to crop specialization. In this context, research that makes it possible to understand land-use changes due to the concentration of agricultural activities is essential. In addition, we can examine adaptive responses to the adverse effects of extreme weather on Brazilian agriculture. Understanding these responses in agricultural adaptation is crucial for prioritizing crop diversification investments and designing risk-mitigating strategies depending on the type of extreme event. The results may support the development of public policies to reduce climate vulnerability through rural extension services or credits.

In this sense, this paper has two main objectives: first, to identify the impact of extreme weather on crop diversification in Brazilian municipalities in the recent past. Events of climate variability include, specifically, frosts in the south region, droughts, heavy rains, and hot days. It uses a fixed-effects panel model for 3818 Brazilian municipalities. The results indicate that increasing consecutive dry days (CDD), dry days, and hot days positively affect crop diversification. This study also understands the effect of future climate change scenarios on the diversification of municipal agricultural production. The results show that the increase in the percentage of hot days in 2045–2065 would increase crop diversification in both scenarios RCP4.5 and RCP8.5 of climate change. The analysis of the impact of extreme climate events on agricultural diversification could help to understand when this practice becomes relevant to be promoted through technical assistance, credit, or research.

The paper is divided into five sections, including this Introduction. The second section presents the empirical strategy, the econometric model, the future simulation specifications, and the data source. The third section presents the empirical results, while Sect. 4 presents a discussion and explores policy implications. Finally, Sect. 5 presents the conclusions.

In this study, we used a panel data model. Data at the Minimum Comparable Area¹ (MCA) level (cross-section) were combined with the years of agricultural censuses 95/96, 2006, and 2017 (time series). In this way, greater precision is obtained in the estimates, the possibility of consistent estimation of models that allow unobserved effects potentially correlated with the regressors, and the possibility of learning more about the dynamics of individual behavior (Cameron and Trivedi 2005). Thus, using panel data made it possible to study the dynamics of crop diversification over the period considered in the study, as well as the effects of climate on these dynamics.

Agricultural diversification at the regional level is an aggregate response to individual farmers' crop choice decisions. This immediate allocation of crops responds to several factors, such as climate variables (Asravor 2018; Rahman 2016). To determine the causal relationship between climate variability and crop diversification, it would be necessary to carry out a natural experiment that would make the farmer's choice of crops a random decision. So, extreme weather can be considered an ideal experiment since climate anomalies cannot be predicted precisely. This can also cause exogenous and random variations in the farmers' decision to allocate crops, affecting crop diversification at the municipal level. Then, it was possible to compare the diversification of the municipalities and to know the effect of extreme events without having problems of selection bias (Piedra-Bonilla et al. 2020a, b). Therefore, the equation of interest was the impact of the extreme weather on crop diversification:

in which vectors of climate variables (C) and control variables (X) affected the Brazilian municipal crop diversification. The climate specifications include five-year moving averages of extreme climate indices for each period of the Agricultural Census 95/96, 2006, and 2017. The five-year moving average was chosen to consider the impact of the medium-term climate on perennial and temporary crops because a more extended period could dilute the effect on these crops (Piedra-Bonilla et al. 2020a, b). This study considers the following extreme events: frosts, droughts, hot days, and extreme precipitation days because they are linked to abiotic stresses in agriculture (Taiz and Zeiger 2006). Thus, five extreme climate indices were used, recommended by the World Meteorological Organization—WMO's Expert Team on Sector-Specific Climate Indices (ET-SCI) for the agriculture sector. The definition of the extreme climate indices for this study is presented in Table 1. These indices were calculated using daily values of precipitation and maximum and minimum temperature with data from the Terrestrial Hydrology Research Group (THRG) (Sheffield et al. 2006), obtaining results of annual values considering the base period (1985–2016) and using standardized software (ClimPACT2) (Alexander & Herold 2015). The base period was considered a minimum period of 30 years that included agricultural census periods.

The daily available temperature and precipitation data were up to 2016. The data source section provides more details.

Regarding the impact of frosts in the South, non-statistically significant parameters can be justified since the diversification strategy is not considered standard practice to manage this type of risk. It would imply, above all, using species or varieties that are tolerant at low temperatures or trees with taller species to help reduce heat-related losses. This would limit the selection of species diversity only for that extreme climatic event. Thus, crop diversification is not the primary strategy to mitigate the risks of frosts. Additionally, we found ambiguous results in nonlinear analyses.

Regarding prolonged droughts (CDD), crop diversification is not considered the main response to the adverse effects of this extreme event in Brazil. Irrigation is the leading practice in reducing the risks of droughts since evidence has responded to the reduction of precipitation in Brazil (Cunha et al. 2015). However, it is worth mentioning that adequate adaptation planning should include synergy between adaptive practices (Teklewold et al. 2017). For example, adopting diversification with irrigation can help overcome prolonged drought shocks. According to Renard and Tilman (2019), the diversity of agricultural species associated with irrigation increases the temporal stability of national crop production, contrary to the instability of precipitation and temperature. Besides, crop diversification is more relevant in areas with less precipitation, such as the Northeast region, unlike the North region, which receives higher rainfall. The Northeast region has approximately 46% of Brazilian farms, mostly family farms (79,2%) (IBGE 2017), and it falls into the highly diversified category (Fig. 2).

Dry days promote crop diversification in Brazil and most regions. However, the results show a negative relation in the Midwest, characterized by a significant area of specialized crops. (Fig. 2). In this region, the most important agricultural practice is the double cropping of soy-maize, a type of diversification. Still, soy-maize cropping is negatively impacted by reduced rainfall during the growing season, making this form of diversification irrelevant in mitigating drought risks (Brumatti et al. 2020). According to Piedra-Bonilla et al. (2020a, b), there is a negative correlation between farm size and crop diversification in Brazil, and the farms in the Midwest tend to be the largest.

The study results indicate that diversification is not used to reduce agricultural vulnerability caused by extreme precipitation days in Brazil. Moreover, diversification may not be influenced by the decreasing trend in days with heavy rains in all Brazilian regions (as shown in Fig. 2.d). However, the agroforestry system, a form of crop diversification, has been recognized as a practice that mitigates the effects of heavy precipitation. Forests are less susceptible to flooding and soil erosion, which reduces water flow (Baldwin et al. 2022). Despite this, the southern region exhibits a significant positive effect. The region has experienced heavy rainfall for about 20 to 30 days from 1987 to 2016, as shown in Fig. 1.c. During the same period, the South had the highest crop diversification level, according to Piedra-Bonilla et al. (2020a, b). Small farms and family farming characterize the region, as IBGE (2017) reported.

Our results suggest that municipalities adopting the diversity of on-farm activities as a strategy for adapting to heat shocks considerably increased on average in Brazil and most regions. These results are similar to those found in the study by Birthal and Hazrana (2019), in which the thermal stress shocks of the previous year increased the diversification of cultures of Indian farmers. Similarly, the results of Dillon et al. (2015) revealed that the number of crop groups harvested has a positive relationship with the degree days shocks.

The channels of extreme weather's impact on crop diversification may take different forms. First, diversification is related to managing risks so that farmers can reduce revenue risk depending on more than a single output (Chavas 2023). Different types of crop diversification are effective technologies for enhancing the resilience of crop yields after a weather-related shock. These technologies include crop rotation (Renwick et al. 2021; Marini et al. 2020), intercropping (Aslam et al. 2020; Amanullah et al. 2021), and livestock-crop systems (Peterson et al. 2020; Delandmeter et al. 2024). Finally, climate change has global agricultural price impacts (Yusifzada 2023), which can increase volatility and encourage farmers and regions to diversify production (Roest et al. 2018).

For future climate change simulations, crop diversification would decrease from 2016 to 2035 (Table 4 and Fig. 4). This would reflect a continuation of the past evolution, decreasing from 1995/96 to 2017. According to Parré and Chagas (2022), the growth rates of agricultural diversification indexes were negative in Brazil between 2002 and 2018. However, in the second period (2046–2065), there would be a change in direction in which diversification would increase in Brazil, albeit slightly (Table 4 and Fig. 4). Then, in this period, crop diversification would start to gain prominence as a strategy for adapting to climate change. These results are similar to those found in Latin America by Seo and Mendelsohn (2008), whose simulations indicate that the increase in crop-livestock integrations would occur in hot, dry, and humid scenarios until 2060.

Furthermore, the increase in the rate variation in diversification in the Midwest in the period 2045–2060 in both scenarios of GHG emissions (RCP 4.5 and RCP8.5) (Fig. 4) shows that this region that has been characterized by the specialization of cultures, such as corn, soy, sugar cane, would change their technologies to more resilient agricultural practices. For example, this region has adopted the system of no-till in grain crops (Romeiro 2014) that requires the rotation of diversified crops for its correct functioning in the long term.

However, the slight growth in crop diversification from the middle of the twenty-first century would not be enough to reduce vulnerability in the face of climatic scenarios. The benefits of resilience acquired from diversification improve in the long term (Bowles et al. 2020; Birthal & Hazrana 2019), so pressing public policies are needed to help increase the diversity of cultures in Brazil to improve its potential to reduce climate risks over time. In this context, it is essential to discuss, on the one hand, the factors that have favored the concentration of crops from the public sector. Still, on the other hand, different current public strategies can encourage agricultural diversification as a Brazilian agricultural adaptation to climate change.

In Brazil, monocultures on large farms destined for export have had economic importance since colonial times, except in the Northeast, because their climatic conditions have been less favorable (Fausto 1994). However, after 1980, the country stopped depending on food imports only after several public policies. Credit, research, and the rural extension of various institutions consolidated agricultural production around the 70 s of the last century (Vieira Filho and Fishlow 2017). However, agricultural technology development was mainly focused on grain production with intensive mechanization, fertilizers, and pesticides, especially in the Southeast and Midwest regions (Alves et al. 2013). This logic was influenced by the practices and technologies of the Green Revolution, which expanded in industrialized countries through the improvement of new varieties of grains, the intensive use of pesticides, mechanization, and irrigation (De Andrades & Ganimi 2007). Thus, the Research and Development (R&D) focused on crop specialization, influencing technical assistance and rural credit equally. So, public policies have promoted this logic of monocultures and a reduction in the diversity of cultures.

However, after pressure from civil society and scientists, R&D has started to focus on developing sustainable technologies, indicating the negative environmental impact, monocultures, and the intensive use of chemical inputs and machinery. In 2009, Brazil established the National Policy on Climate Change, in which the agricultural sector is contemplated through the Low Carbon Emission Agriculture Plan (ABC Plan), established in the following year until 2020. The ABC Plan has organized actions to adopt sustainable production technologies to meet the country's GHG emissions reduction commitments in the agricultural sector. Within the actions, the concept of diversification is not considered as such. Still, several forms of diversification are used in the programs, such as Crop-Livestock-Forest Integration, Agroforestry Systems, and rotation, intercropping, or crop succession as part of the No-Tillage Systems. According to Manzatto et al. (2018), it is observed that the ABC Plan has driven agricultural diversification through its programs, even exceeding the commitments established at the beginning of the programs, since the Crop-Livestock-Forest Integration reached 146% of its goals. So, the ABC Plan must continue in a new phase to reinforce and expand resilient and mitigation technologies, such as diversification. According to Souza Piao et al. (2021) and Vinholis et al. (2021), to strengthen the adoption of the ABC Plan technologies, it is necessary to improve the rural extension service in the diffusion of sustainable technologies to producers. These results accompany the positive effects of technical assistance on agricultural diversification observed in most climate specifications.

On the other hand, remembering that smaller farms are more willing to diversify their agricultural production in Brazil, discussing public policies focused on family farming is important. Firstly, there are public food purchases, as instruments to encourage the purchase of products from family farming, such as the Food Acquisition Program (Programa de Aquisação de Alimentos—PAA), Institutional Purchases, and the National School Feeding Program (Programa Nacional de Alimentação Escolar—PNAE). These programs aim to support sustainable agricultural production and the acquisition of diversified foods considered to be of food value (Grisa et al. 2020). Therefore, public policies have contributed to the increase in the sale of diversified products in family farming.

Here, we show the impact of extreme weather events on crop diversification in Brazil for the first time. First, the results showed that Brazilian municipalities adopted diversification as an adaptation strategy when the climatic shocks were dry spells (CDD), days without precipitation (Rn1mm), and hot days (TX90p). Notably, the evolution of the percentage of hot days and the average annual temperature had increasing trends from 1985 to 2016. In the Southern region, diversification was not considered a practice to mitigate the vulnerability caused by frosts and heavy rainfall in the country. However, the heavy rainy days decreased in all Brazilian regions over the same period, while the frosts were erratic. Thus, the response to diversification as an adaptive strategy would depend on the type of climate shock and the characteristics of the climate and agricultural technologies of the regions. Each region's specific models show that extreme climate indices' impact on crop diversification varies across the Brazilian regions, so other adaptative strategies should accompany it.

The results of simulations of the rate variation in Brazilian crop diversification in the beginning (2016–2035) and mid (2046–2065) of the twenty-first century show the following forecast. In the first period, all-climate specifications showed a decrease in the percentage variation of Brazilian agricultural diversification, especially in the RCP8.5 scenarios. However, the decrease in Brazilian diversification in 2016–2035 would be less than the historical average observed (−8.7%) from 1996 to 2017. Next, the variation in diversification would increase in the scenario of high GHG emissions (RCP8.5) in the second period (2046–2065). These forecasts of changes in agricultural land use in Brazil indicate convergence to agricultural systems that are more resilient to climate change.

The findings of this study must be considered in light of some limitations. First, a change in diversification at municipality levels may not necessarily imply a similar change in the degree of diversification at the scale of farms. Nevertheless, it is assumed that the contribution of individual activities impacts diversification with data at the municipality level. Furthermore, the municipality analyses help focus on providing credit and technical assistance through administrative levels (Birthal & Hazrana 2019). Second, it was not possible to empirically estimate which form of diversification would be projected in the future. But discusses the concentration of on-farm products that could increase with the increase of some extreme weather events. We recommend that upcoming studies take into consideration these concerns.

Finally, promoting sustainable agricultural practices, such as crop diversification, is essential through extension and rural credit. However, based on the research results discussed above, it can be concluded that crop diversification is not a one-size-fits-all solution to mitigate the risks of extreme weather events in Brazil and should not be the only adaptive practice adopted. Instead, it should be complemented with other adaptive strategies such as irrigation and a no-till system. Public policies should focus on providing support and incentives for farmers to adopt these practices and invest in research and development to identify new and more effective strategies to cope with climate change. Furthermore, policymakers should also consider the regional differences in Brazil and tailor their policies accordingly to meet the specific challenges faced by farmers in each region.