Predictive Threshold Modeling of Evolutionary Plasticity and Colony Collapse Risk in Apis Mellifera Through Climate-Driven Architectural Degradation, BehavioralEnergetics, And Material Economy Using YOLO-Seg, and Sentinel-3 Framework

Abstract Climate-induced change now stands out as a critical driver of how social insects adjust their behavior and build nests - though exactly how temperature links to nest architecture still lacks clear quantification. A method combining artificial intelligence with satellite data helps track niche architectural drift in Apis mellifera , focusing on the degradation of hexagonal comb symmetry when the weather turns unstable. Satellite records from Sentinel-3, along with MODIS and ERA5-Land systems, feed into a detailed climate picture spanning 24 years, stretching back to 1999 and reaching toward 2026. Google Earth Engine processes these inputs across three areas marked by different ecosystems: one in Canada near Prince Albert at 53.13°N and 106.35°W, another in Morocco’s Rabat region located at 33.97°N and 6.84°W, plus a third site situated eastward in Lithuania’s capital zone, Vilnius, around 54.68°N and 25.27°E. To evaluate high-resolution comb images from detailed comb scans, a Symmetry Coefficient (S c ) emerged using YOLOv11-seg for instance segmentation alongside OpenCV-driven symmetry calculations. Instance segmentation mapped each cell position, and the behavior patterns quantification followed through pose estimation models like OpenBeePose and VGG-19. Instead of relying on surface metrics, material flow was assessed via Wax-to-Storage Ratio (WSR), built from 3D reconstructions to evaluate material efficiency. Data point to one clear trend: when heat fluctuation rises, hive structure weakens significantly (β = −0.08495, p < 0.01). Below a critical symmetry threshold, S c value of 0.6214, survival odds drop sharply, with collapse probability reaching 70%. As temperatures climb, information in waggle signals loses coherence; entropy measures expose systemic breakdowns deeper than mere shape changes. By this comparative study, differential phenotypic plasticity is confirmed, which shows how Boreal groups handle change compared to Mediterranean ones - their stability score sits at 0.852, noticeably higher resilience than the 0.665 found southward. The research develops a predictive ecological threshold paradigm that connects behavior, energetics, environmental stress, and structural effects. Instead of just measuring shifts, it builds bridges between machine learning and real-world animal actions amid warming climates. Early warnings for bee network instability become possible through this lens, scaling across regions without losing precision.