1993—2017年气候变化对西藏谷物单产的定量影响

气候变化对农业的影响是全球关注的热点问题之一,青藏高原对气候变化尤其敏感,但气候变化对青藏高原农业产量的定量影响缺乏系统研究。为定量评估气候变化对西藏谷物单产的影响,本文使用气象数据与年鉴统计数据,选取了固定效应模型、差分模型和线性去趋势模型3类统计模型,分析了1993—2017年间气候变化（最低气温、降水量、生长度日和太阳辐射）对西藏县（区）级、市级和自治区3个尺度的谷物单产的影响。结果表明：西藏整体对于温度（最低气温和生长度日）的敏感性大于降水量和太阳辐射。各项气候因子对西藏谷物单产的整体影响为正影响,但不同区域对气候因子的敏感程度和显著性不同。除了生长度日对于拉萨为负影响以外,最低气温、降水量和太阳辐射对于所有市均为正影响。气候趋势对于西藏整体谷物单产的影响为正影响,不同模型计算结果集中在1.5%~4.8%区间内。3类模型中固定效应模型稳定性最好,线性去趋势模型好于差分模型,差分模型在引入气候因子间的交互项后模型稳定性降低。本文有助于西藏实施更加有空间针对性的农业适应气候变化措施,以应对气候造成的青藏高原农业生态系统变化。

Quantitative evaluation of the effects of climate change on cereal yields of Tibet during 1993-2017

Climate change, which can lead to environmental problems, has become a global concern, especially in the Tibetan Plateau. However, there are few studies on quantitative assessment of climate change on agricultural crops in the Tibetan Plateau. We used meteorological data and yearbook statistical data to assess the impact of climate change on cereal yields of Tibet. Three types of statistical models were selected, including fixed effect model, first-difference model and linear de-trend model. We analyzed the impacts of climate change (minimum temperature, precipitation, growing degree days, and solar radiation) on cereal yields at the county, city, and autonomous region scales in Tibet from 1993 to 2017. The results showed that the sensitivity of cereal production to temperature (minimum temperature and growing degree days) was greater than that to precipitation and solar radiation. The joint impacts of climate variables were positive, but the sensitivity and significance of different regions were different. Except for the negative impact of growing degree days on Lhasa, the impacts of minimum temperature, precipitation and solar radiation were positive on all cities. The impacts of climate trends on the cereal yields in Tibet Autonomous Region were positive and the results of different models were concentrated in the range of 1.5% to 4.8%. Among the three types of models, the fixed effect model performed best in robustness, and the linear de-trend model was better than the first-difference model. After we add the interaction between different climate variables, the robustness of the first-difference model decreased. Our study could help implement more spatially targeted agricultural adaptation measures to cope with the impact of climate change on the agricultural ecosystem of the Tibetan Plateau.