The creation of a satellite rapid alert programme (DETER-A) in 2004 was a cornerstone of the Brazilian government’s strategy to reduce deforestation. This programme allowed authorities to detect and respond rapidly to periods of deforestation. Due to the fact that the policy instrument was based on multispectral remote radar, weather-related obstacles posed a continuous impediment to the study of deforestation. This paper investigates to what extent cloud cover has reduced the effectiveness of the DETER-A program to detect deforestation. To test this hypothesis, survival model analysis is undertaken on satellite data derived measures of local deforestation. The emphasis is on the state of Maranhão, which is separated into two areas by an arbitrary line of demarcation (Legal Amazon delimitation) where the forest on one side is covered by the satellite monitoring program while the other is not. The results suggest that following the implementation of the satellite monitoring program, there was more deforestation in those years with more cloud persistence in the area covered by the program. Counterfactual simulations indicate that the absence of clouds would have prevented deforestation equivalent to almost 7% of the study region, which is equivalent to 73 million tonnes of CO2 with a value of US$ 366 million. If the current monitoring system was replaced with the experimental policy instrument DETER-C/DETER INTENSO, cloud cover would be less of an impediment to deforestation detection.