A Diachronic Geospatial Analysis of Cochineal-Induced Degradation of Prickly Pear (Opuntia spp.) in the Rhamna Province (Morocco): An Innovative Framework Using Remote Sensing and GIS

Abstract

Prickly pear (Opuntia spp.) is more than a conventional crop in Morocco; it serves as a keystone species sustaining rural livelihoods and enhancing ecosystem resilience. Owing to its remarkable tolerance to drought, poor soils, and heat stress, Opuntia has become integral to sustainable agriculture, land restoration, and national programs such as the Green Morocco Plan aimed at combating desertification. However, this success has been critically threatened by the rapid spread of the cochineal (Dactylopius opuntiae), a devastating pest that has decimated cactus plantations nationwide. To quantify the magnitude and spatial extent of this invasion, we performed a high-resolution geospatial analysis in Rhamna Province—one of Morocco’s main cactus-producing regions—between 2014 and 2020. A grid-based visual interpretation method was applied to sub-meter Google Earth imagery, supported by 3,305 ground control points and GIS post-processing. Cactus cover declined sharply from 23,326 ha in 2014 to only 267 ha in 2020 (−98.9%), with remaining stands highly fragmented across roughly ten communes. The corresponding annual revenue loss, estimated from fruit and biomass conversion factors and farm-gate prices, reached approximately 230.6 million MAD (≈ 23.8 million USD). Accuracy assessment based on confusion matrices yielded high overall agreement (overall accuracy and Cohen’s κ with 95% confidence intervals), confirming the reliability of the results. These spatially explicit, accuracy-controlled findings reveal the dramatic contraction of Opuntia cultivation and provide a quantitative baseline for recovery monitoring. The study highlights the urgent need for coordinated pest management and deployment of resistant genotypes, while demonstrating the value of expert-guided visual interpretation for monitoring bio-invasions in fragmented arid agroecosystems where conventional spectral classification often fails to capture fine-scale dynamics.