The relationship between climate change and conflict is a significant area of research and policy concern. While weather extremes, particularly meteorological droughts, have been linked to armed conflict outbreaks, the causal mechanisms remain debated. The most commonly proposed mechanism is the impact of agricultural production shocks, leading to migration and ultimately conflict. However, a lack of reliable sub-national data on agriculture in high-conflict regions has hampered rigorous investigation of this hypothesis. Many studies rely solely on meteorological data, assuming a direct correlation between weather patterns and widespread crop failure. This has contributed to a lack of consensus regarding the climate-conflict link. The Syrian conflict has often been cited as a prime example of drought-induced conflict. The severe drought between 2007 and 2009 has been linked to the 2011 uprising and subsequent civil war. The narrative suggests the drought caused an agricultural collapse, leading to mass rural-urban migration, which overwhelmed Syrian cities and contributed to the conflict. While this narrative is plausible, empirical evidence supporting a direct causal chain is limited, relying primarily on anecdotal evidence and temporal correlations. Several alternative explanations for the Syrian conflict exist, including socioeconomic factors like poverty, unemployment, corruption, repression, and a growing rural-urban divide. Even if the drought was a contributing factor, its precise role and the nature of its impact on the agricultural system require further investigation. The existing literature lacks detailed data on agricultural productivity and land-use dynamics to adequately assess whether the drought caused a true agricultural collapse or a temporary shock from which the system recovered. The study will attempt to fill this crucial gap in the literature using satellite data of cropland dynamics in Syria during the pre- and early-conflict period. Therefore, this research aims to answer: How did the 2007-2009 drought affect the Syrian agricultural system before the outbreak of the Syrian civil war in 2011? This overarching question is addressed through three sub-questions exploring climate variability's effects on Syrian agriculture, changes in the agricultural system’s vulnerability to drought between 2000 and 2016, and whether the observed cropland dynamics align with the criteria for an agricultural collapse. The term 'dryness' is used to describe combined meteorological conditions, while 'drought' refers to specific drought periods.

The existing literature on climate change and conflict presents a mixed picture, with studies yielding diverse conclusions on the strength and nature of the relationship. Some research suggests a clear link between climate variability, particularly droughts, and increased conflict risk, often mediated through agricultural production shocks and subsequent migration. Other studies, however, have questioned the direct causal link, highlighting the complexities of the relationship and the influence of various socio-political factors. Furthermore, limitations in data availability, particularly regarding sub-national agricultural data in conflict-prone areas, have hampered rigorous empirical analysis. Regarding the Syrian case, multiple studies and media reports have emphasized the role of the 2007-2009 drought as a trigger or significant contributor to the 2011 uprising and ensuing civil war. These narratives commonly posit an agricultural collapse induced by the drought, resulting in mass rural-urban migration and social unrest. However, this view has been challenged by others who point to the multifaceted nature of the conflict, highlighting the role of political repression, socioeconomic inequalities, and long-standing grievances. The debate highlights the importance of analyzing agricultural impacts not just through meteorological data but also through direct observation of agricultural activity and its socioeconomic implications. While some studies attempted to link meteorological drought to conflict in Syria, they generally lack detailed, granular agricultural data, leading to varying interpretations of the evidence. This paper seeks to address these gaps by leveraging satellite-based data to directly assess changes in agricultural activity, providing a more comprehensive understanding of the complex relationship between drought, agriculture, migration, and conflict in the Syrian context.

This study uses a combination of satellite-based remote sensing data and statistical modeling to analyze the impact of the 2007-2009 drought on the Syrian agricultural system. The core data source is a land-use/land-cover dataset, covering Syria and Iraq, that provides annual cropland extent (single and double cropping) from 2000 to 2016. This dataset was generated using 8-day NDVI time series from MODIS Terra data, processed to remove cloud contamination and smoothed using a 30-day moving average. A two-stage random forest classification approach, trained on high-quality Landsat data, was used to map cropland extent, using spectral-temporal metrics derived from the MODIS NDVI time series. To quantify fallow areas, a 'normal cropland extent' was defined based on the most frequent cropland class (mode) over the study period. Fallowness was then calculated as the difference between the normal cropland extent and the actively cropped land each year. A rigorous accuracy assessment was conducted using independent validation points, achieving an average overall accuracy of 90% for the annual fallow/cropland maps. The accuracy assessment involved visually interpreting MODIS phenological evolution and confirming interpretations using Google Earth imagery. This process provided error matrices for each year, enabling the calculation of overall classification accuracy, user's and producer's accuracies, and confidence intervals around area estimates. The statistical analysis focused on investigating the relationship between dryness (measured by the 6-month Standardized Precipitation-Evaporation Index, SPEI6) and fallowness. Spatial fixed-effect models were employed to account for time-invariant heterogeneity and spatial dependence, using the spatial Durbin model with a maximum likelihood estimator. This approach considers the impact of dryness on fallowness, controlling for the potential for farmers in neighboring areas to influence each other's cropping decisions. The models were run for different time periods (pre-drought, post-drought, conflict years) and for different sub-regions of Syria, including the northeastern governorates (particularly affected by the drought and government policies) and areas with differing levels of irrigation dependence. To assess potential increases in agricultural vulnerability after the drought, the models analyzed the changes in the sensitivity of fallowness to dryness over time. Additional analyses examined the impact of local violence and the removal of government fuel and fertilizer subsidies. The study uses the concept of regime shifts in social-ecological systems to rigorously assess if an agricultural collapse occurred based on four criteria: abruptness of change, substantial loss, persistence of loss, and structural change. The study also considered national-level livestock data to validate the findings on cropland dynamics.

The study's key findings challenge the prevailing narrative of a drought-induced agricultural collapse in Syria preceding the 2011 conflict. Analysis of satellite-derived cropland data revealed substantial year-to-year variations in fallowness, with the highest levels (50%) observed in 2008 at the peak of the drought. However, a significant finding was the rapid recovery of cropland in the following years, with near-full recovery by 2010. This rapid rebound contradicts the assertion of a sustained agricultural collapse. Only 0.5% of normally active cropland showed signs of permanent abandonment after 2008. National-level livestock data confirmed the pattern observed for croplands, indicating a decline during the drought and subsequent recovery. Statistical modeling demonstrated a strong correlation between dryness (SPEI6) and fallowness, with drier periods associated with increased fallowness across the entire study period. However, the analysis revealed an increase in drought sensitivity in the post-drought period (2008-2016), especially in the conflict years (2013-2016). This heightened sensitivity cannot be fully attributed to irrigation dependence or local violence, suggesting underlying systemic changes may have contributed to increased vulnerability. While irrigated areas did not show significantly increased vulnerability after subsidy cuts, both irrigated and rain-fed areas experienced similar effects, with areas having below-median irrigation levels showing slightly higher sensitivity to dryness. The analysis of the northeastern governorates showed similar patterns to the national-level results. Considering the four criteria for agricultural collapse (abruptness, substantial loss, persistence, and structural change), the findings do not support the collapse narrative. While the drought-induced reduction in active cropland was abrupt and represented a substantial, though temporary, loss, the rapid recovery and lack of persistent land abandonment refute the criteria for a collapse. While the agricultural system showed signs of increasing vulnerability to drought, this was likely influenced more by factors like subsidy cuts, unsustainable agricultural practices, and, to a lesser extent, local violence, than by the drought itself. The study notes that the analysis focuses on cropland and does not fully encompass all aspects of the agricultural system. Yet, the pattern of temporary decline and subsequent recovery challenges the narrative of permanent rural-urban migration due to agricultural collapse.

The findings of this study directly challenge the widely held belief of a drought-induced agricultural collapse in Syria that significantly contributed to the 2011 conflict. The rapid recovery of cropland after the 2007-2009 drought suggests the agricultural system possessed significant resilience and capacity to recover from such shocks. The observed increase in drought vulnerability in later years highlights the importance of considering long-term factors, such as unsustainable practices and policy changes, in shaping the resilience of agricultural systems. The study underscores the need for a nuanced understanding of the climate-conflict nexus, moving beyond simple correlations between meteorological droughts and conflict outcomes. The study’s methodology, focusing on high-resolution satellite data of cropland dynamics, offers a significant improvement over previous research that often relied solely on meteorological data. This approach provides a more direct assessment of agricultural impacts, allowing for a more accurate evaluation of the drought’s consequences. While the study primarily focused on cropland, the consistent pattern with national-level livestock data adds to the robustness of the findings. While the study suggests that temporary migration may have occurred during the drought years, it refutes the narrative of a mass and permanent rural-urban migration. This study contributes to a broader understanding of the complex interplay between climate change, agriculture, and conflict, challenging overly simplistic narratives that emphasize climate as the sole or primary driver of conflict. The findings have implications for the development of climate-resilient agricultural policies and strategies, highlighting the need to consider long-term systemic vulnerabilities in addition to the immediate impacts of weather events.

This study provides strong evidence against the hypothesis of a drought-induced agricultural collapse in Syria prior to the 2011 conflict. The findings highlight the resilience of the Syrian agricultural system and the importance of considering a range of socio-political and economic factors alongside climate variability in understanding conflict dynamics. The use of high-resolution satellite data offers a more nuanced perspective on the relationship between meteorological drought and agricultural outcomes than previous research, which primarily relied on broad-scale meteorological data. Future research should explore further the interactions between climate variability, agricultural policies, socio-economic factors, and conflict, using similarly high-resolution data across various regions and contexts.

While the study uses high-resolution satellite data to capture cropland dynamics, it acknowledges that this data does not encompass all aspects of the agricultural system. For example, the study does not explicitly measure livestock production or the impact of drought on other agricultural products. Furthermore, the statistical analysis relies on publicly available data on violence, which might not fully capture the complex reality of ground-level conflict and its impact on agricultural activities. The analysis also focuses primarily on cropland, which is one aspect of the agricultural system. Finally, the study implicitly assumes a correlation between land fallowness and reduced agricultural production, which, while likely, is not empirically confirmed with additional data. Despite these limitations, the study's findings significantly contribute to a better understanding of the complex climate-conflict relationship in Syria.