Abstract
Climate change affects agriculture worldwide, with stronger socio-economic impacts in low-income countries where the lack of data hinders the implementation of effective interventions. This paper proposes an approach to assess local effects associated with climate change in data-scarce contexts by integrating farmers’ perceptions with available climate data. Tested in the Upper Huallaga basin of the Peruvian Amazon, the study combines daily weather station data and ERA-5 reanalysis with 73 structured interviews. Results show an increasing temperature trend of 0.2 °C per decade, consistent with farmer perceptions. While farmers also reported increased wind gusts and precipitation—contrary to daily quantitative records—analysis of annual crop water deficit and surplus volumes suggests an increase in sub-daily convective rainfall events, explaining this discrepancy. This integrated approach effectively assesses climatic alterations and locally driven adaptation measures in data-scarce regions.
Introduction
Smallholder farmers in low-income countries are uniquely vulnerable to climate shifts, as they rely heavily on rainfed agriculture. Developing effective adaptation strategies requires high-resolution local climate data, which is often missing or unreliable in these regions. Traditional monitoring is costly, and global models frequently fail to capture micro-climatic nuances in complex terrains like the Amazon. This study addresses the challenge of assessing climate impacts in "data-scarce" contexts, specifically the Upper Huallaga Valley in Peru, by bridging the gap between instrumental records and local ecological knowledge.
Methodology and Study Area
The researchers employed an Integrated Assessment (IA) approach, synthesizing quantitative meteorological data with qualitative social surveys.
Quantitative Data
The study utilized historical daily and monthly records (1940–2019) from the Tingo María weather station. To ensure data quality, the Standard Normal Homogeneity Test (SNHT) was applied, and ERA5 reanalysis data was used to fill gaps. A critical component of the analysis involved calculating reference evapotranspiration and crop water deficit/surplus volumes to translate raw weather data into agricultural impacts.
Qualitative Data
Structured, face-to-face interviews were conducted with 73 banana farmers in the Huallaga River floodplains. These surveys captured long-term perceptions of climate change, observations of extreme weather, and current adaptation strategies.
Key Results and Insights
The Perception-Data Mismatch
A significant finding was the divergence between daily instrumental records and farmer observations regarding precipitation and wind. While quantitative models showed no statistically significant multi-decadal trends for these variables, over 80% of farmers perceived changes in precipitation, and 73% noted an increase in wind gusts.
The researchers hypothesized that daily data masks sub-daily extreme events. By modeling crop water deficit, they found a recent trend of simultaneous increases in water surplus and dry months. This suggests that rainfall is occurring in shorter, more violent convective bursts—leading to immediate runoff rather than soil infiltration—which aligns perfectly with the farmers' reports of increased "heavy rain" despite stable annual totals.
Temperature and Heat Extremes
Both homogenized station data and ERA5 confirmed a warming trend of approximately 0.2 °C per decade. This trend was accurately perceived by 70.2% of the surveyed farmers. Furthermore, the frequency of extreme heat days (exceeding the 99th percentile) has risen sharply since 2010.
Socio-Political Barriers to Adaptation
Despite high awareness of climate change, 53.3% of farmers had not adopted adaptation measures. The study identifies a deep-seated mistrust of institutions, rooted in the region's history of illegal coca cultivation and subsequent conflict, as a primary barrier. Adaptation is thus revealed to be a sociological challenge as much as a technical one.
Conclusion
Integrating local perceptions provides a "high-resolution sensor network" that can fill critical gaps in instrumental records, particularly regarding sub-daily extreme events. However, the study also highlights the urgent need for improved meteorological networks capable of sub-hourly recording in data-scarce regions. Future adaptation initiatives must look beyond technical solutions and account for local socio-political contexts to successfully implement grassroots strategies like reforestation.
Summary updated