基于RAGA优化投影寻踪模型的黑龙江省农业旱灾风险评价

Comprehensive Evaluation of Agriculture Drought Risk in Heilongjiang Province Based on Projection Pursuit Model Optimized by Real Coding Based Accelerating Genetic Algorithm

基于投影降维思想,结合黑龙江省农业生产实际状况构建农业旱灾风险评价指标体系,采用基于实数编码加速遗传算法优化求解最佳投影方向,从时空两个角度评价黑龙江省农业旱灾风险。结果表明:2006—2015年黑龙江省农业旱灾风险呈现下降的时序特征,且主要受干旱危险性和地区抗旱能力两个要素影响。危险性变化具有波动性,暴露性缓慢增加,抗旱能力呈现稳定提高的特征;2015年黑龙江省农业旱灾风险呈现西部地区明显高于东部地区的空间特征,主要由西部地区高危险性、高暴露性、低抗旱能力导致。处于旱灾高风险区的是齐齐哈尔、绥化、黑河三市,处于旱灾低风险区的是鹤岗、伊春、牡丹江三市。

To comprehensively evaluate agriculture drought risk in Heilongjiang Province from the perspective of time and space, the projection pursuit model optimized by real coding based accelerating genetic algorithm(RAGA) was applied and the agricultural drought risk evaluation index system was constructed. The temporal variation results indicated that the agricultural drought risk in Heilongjiang Province decreased from 2006 to2015, mainly affected by the drought dangerousness and the resistance capacity. The dangerousness had volatility, and slowly increased exposure, and the resistance capacity showed increased stability. The spatial distribution results indicated that the drought risk in the western regions was significantly higher than that in the eastern regions. The differences were primarily caused by higher drought dangerousness, higher exposure and lower resistance capacity in the western regions of Heilongjiang Province. The results also showed that Qigihar, Suihua and Heihe City were located in high-risk areas and Hegang, Yichun and Mudanjiang were located in low-risk areas.