粮价是气候变化影响向社会传递过程中的重要环节之一。中国历史时期气候变化、粮价及社会影响的相互关系研究，有助于完善对中国历史气候变化影响机制与关键过程的认识，可为充分认识中国当前粮食安全可能面临的气候变化风险提供历史参考，有助于深入理解中国当前的粮食安全战略，更有力地保障国家粮食安全。

清代处在小冰期气候及向现代暖期过渡的阶段，在此期间华北地区气候的阶段性冷暖变化及旱涝灾害变化对社会经济产生了显著影响，并导致社会不稳定性增加。清代华北地区气候变化重建，及其对农业生产、人口、社会系统的影响研究已经取得大量成果，但粮价与气候变化的关系及其在气候变化影响传递过程中的作用尚较少涉及。

本文收集、整理清代粮价及货币物价、气候及旱涝灾害，以及农业生产系统、经济系统、人口系统及社会系统相关数据资料，通过系统的数据处理过程，获得1736-1911年华北地区的府、分区、全区等不同空间尺度的年分辨率小麦原始粮价及去趋势粮价数据。在此基础上，采用描述性统计、相关分析、小波分析等统计分析方法，以及空间分析与制图方法，对清代华北地区气候变化与粮价波动的关系进行了探讨。主要工作包括：

（1）揭示了清代华北地区粮价的时间变化特征、空间差异及空间关联特征，以及时空变化的综合特征。

（2）从时间序列的阶段性变化、周期变化的对应关系及相关关系，时空变化格局的关联等角度，对清代华北地区气候变化与粮价波动的关系进行了定量分析。

（3）辨识清代华北地区粮价波动的气候变化敏感因子，分析气候变化影响粮价波动的可能路径。揭示清代华北地区粮价气候变化敏感性的时空差异，并探讨了农业生产条件、水路运输条件、耕地承载力、经济子系统弹性等因素对粮价气候变化敏感性的可能影响。

（4）定量分析清代华北地区粮价波动与饥荒、疫灾、动乱等人口和社会危机，以及赈贷、蠲缓等社会响应行为的关系。

得到的主要结论如下：

（1）清代华北地区年降水量、旱灾强度及范围与去趋势粮价的对应关系存在明显的阶段性。1781-1890年的气候寒冷期，年降水量、旱灾强度及范围与去趋势粮价的对应关系明显增强。期间华北全区及各分区1782-1792年、1808-1817年、1831-1838年及1873-1880年等持续少雨期，旱灾强度及范围相对偏高，多对应去趋势粮价高峰期或急剧上升时段。85.1%的区域-时段组合的旱灾强度或范围指数与去趋势粮价具有良好对应关系，其中持续少雨期中该比例则达96.9%。

（2）旱灾对年际粮价波动的影响更为显著，旱灾指标与去趋势粮价存在显著的相关关系和周期共振，且具有明显的时间滞后和空间波及效应。华北全区旱灾强度、范围指数与去趋势粮价正相关（r分别为0.303、0.337）的显著性明显高于降水，粮价滞后1年时相关关系显著性最高（r分别为0.545、0.486），其影响至少可持续2年。1781-1890年去趋势粮价在18-30年、8-12年的主导周期上与旱灾强度及范围均存在显著的正相位周期共振现象，在8-12年的短周期更为显著。全区旱灾强度、范围指数与各分区去趋势粮价也存在具有滞后效应的正相关和正相位周期共振。

（3）粮价气候变化敏感性存在显著的阶段性变化，受到气候冷暖和干湿变化的联合效应，以及耕地承载力、经济子系统弹性等因素的影响。以相关系数、方差解释率等衡量的年际粮价波动对各气候变化因子的敏感性，在1770s-1830s、1870s-1890s显著增加。这些寒冷时段温度对粮食生产的不利影响，使降水、旱灾等因子对粮价波动的影响更为显著。人均耕地面积在1880年之前持续降低并于1780s接近温饱阈值，期间多数时段（除1841-1865年）旱灾强度及范围指数与去趋势粮价的相关系数及方差解释率均达到极显著水平（P<0.001）。清政府粮食储备大幅降低、财政平衡恶化的1796-1815年、1856-1880年，粮价气候变化敏感性显著增强，而1736-1770年、1816-1855年、1881-1911年等经济子系统弹性有所增强时段的敏感性则明显降低。

（4）粮价气候变化敏感性具有明显的区域分异和空间聚集特征，主要受到农业生产条件和水路运输条件的显著影响。华北地区去趋势粮价指数与不同空间尺度气候变化因子的相关性及最优回归模型的方差解释率，均呈现以山西省南部、直隶省南部及河南省西北部、西部等高值区为中心，向周边区域递减的空间分布格局。黄土高原区、黄-海平原区、黄河下游区、淮北-鲁东区等不同农业气候区的粮价气候变化敏感性各项指标均存在显著差异（P<0.001），其中粮价对局地气候变化因子敏感性区域差异的显著性更高。水路运输条件作为粮食市场关联的重要因素，可显著影响粮价对全区气候变化因子的敏感性。有水路通过的府（州），与其他区域粮价差分相关系数均值显著偏高（P=0.004），而相关系数均值与去趋势粮价对全区气候变化因子的敏感性指标具有显著的相关性（r=0.726-0.860）。

（5）清代华北地区去趋势粮价与饥荒的阶段性变化、周期波动及时空变化格局均存在明显的对应关系，两者之间存在显著的正相关及交互影响；不考虑战乱及大范围流行等因素，粮价波动是清代华北地区疫灾的重要影响因素，其影响存在明显的阈值效应。去趋势粮价偏高或急剧上升时段，饥荒县次往往高于全时段平均值。华北地区及各分区去趋势粮价指数与饥荒县次具有18-30年、8-12年的周期共振，呈现稳定的正相位关系，且在1781-1890年的寒冷期具有更高的显著性。多数全区性及局地大范围的粮价指数异常高值，均与饥荒范围指数的高值区具有较强的对应关系。华北全区饥荒指数与当年、滞后1年、滞后2年的去趋势粮价相关系数分别达到0.420、0.641、0.340（P<0.001），在消除粮价自相关的影响后，饥荒指数对去趋势粮价的滞后影响仍然显著。这表明短期粮价异常波动可能引发饥荒，而饥荒通过社会经济系统的反馈作用对粮价波动产生滞后影响。如果不考虑1820-1822年、1862-1864年、1888-1895年、1902年、1861-1868年等时段战乱及外部输入引发的疫灾，去趋势粮价与疫灾指数的相关系数为0.592（P<0.001），超过一定限值时其对疫灾指数的影响显著增强（r=0.699）。

Grain price is one of the important links in the transmission of the impact of climate change on society. The research on the interrelation of climate change, grain price and social impact, can  not only help to improve the understanding of the impact mechanism and key process of climate change in historical China, but also provide historical background for fully understanding the possible climate change risks that China's food security may face at present. Which is therefore beneficial for in-depth understanding of China's current food security strategy, so as to ensure national food security more effectively. 

Climate in the Qing Dynasty was in the transition stage from the Little Ice Age to the Modern Warm Period, which had a significant impact on the social economy and leads to increased social instability. The researches on reconstruction of climate change in North China in the Qing Dynasty and its impact on agricultural production, population and social system had made great achievements. While the relationship between grain price and climate change, and its role in the transmission process of climate change impact is seldom involved.

In this study, we collected and extracted the data of grain and currency prices, climate, drought and flood disasters in the Qing Dynasty, as well as the related data of agricultural production, population, economic and social systems. Then we obtained the data of raw and detrended grain price at different spatial scales in North China in1736-1911, through a series of processing procedures. Based on these data, we discussed the relationship between climate change and grain price volatility in North China in the Qing Dynasty, by applying descriptive statistics, correlation analysis, wavelet analysis and other statistical analysis methods, as well as spatial analysis and mapping methods. The main research work includes:

(1) We revealed the characteristics of temporal change, spatial difference and correlation, as well as the pattern of temporal-spatial changes of grain prices in North China in the Qing Dynasty.

(2) We made a quantitative analysis on the relationship between climate change and grain price volatility in North China in the Qing Dynasty, from the perspectives of the corresponding relation between the phased and periodic change of time series, correlation relation, as well as the links between patterns of temporal-spatial changes.

(3) The sensitive factors of climate change on grain price in North China in the Qing Dynasty were identified and the possible pathways for grain price affected by climate change were analyzed. Then we revealed temporal and spatial differences of the sensitivity of grain price to climate change in North China in the Qing Dynasty, and discussed the possible influence of agricultural production conditions, water transport conditions, cropland carrying capacity, economic system elasticity.

(4) We made a quantitative analysis on the relationship between grain price volatility and population and social crisis such as famine, epidemic disaster and turmoil in North China in the Qing Dynasty, as well as the social response behaviors such as food relief and loan, tax exemption and suspension.

The main conclusions are as below:

(1) There was a significant correspondence between the phased changes of precipitation and drought disaster, and the grain price volatility in North China in the Qing Dynasty. The correspondence between annual precipitation, drought intensity and coverage, and the de-trend grain price became more obvious during the cold period 1781-1890. The drought intensity and coverage were relative high in North China during the periods with continuous low rainfall, such as 1782-1792, 1808- 1817, 1831-1838 and 1873-1880, which mostly corresponded to the peaks or sharp rises of detrended grain price. In 85.1% of the region-periods, the drought intensity or coverage index  had a good correspondence with the detrended grain price, while the proportion reached 96.9% during the aforementioned continuous low rainfall periods.

(2) Drought had more significant impact on the inter-annual fluctuation of grain price in North China in the Qing Dynasty. The drought indexes had significant correlation and periodic resonance with the detrended grain price, with obvious lagged effect and spatial influence effect. The positive correlation between drought intensity and coverage index and detrended grain price (r = 0.303 and 0.337, respectively) in North China was significantly higher than that of precipitation, and the correlation was the highest between the drought indexes and grain price with 1 year lag (r = 0.545 and 0.486, respectively) . In 1781-1890, the detrended grain price had significant periodic resonance with the drought intensity and coverage indexes in the dominant cycle of 18-30 years and 8-12 years. The periodic resonance was at positive phase, with higher significance especially in short cycle of 8-12 years. There were also positive correlation and positive phase periodic resonance between the drought intensity and coverage indexes of the whole North China and the detrended grain prices in different sub-regions.

(3) The sensitivity of grain price to climate change in North China in the Qing Dynasty had significant phased changes, which were affected by the combined effects of temperature and humidity, as well as farmland bearing capacity, economic subsystem elasticity and other factors. The sensitivity of inter-annual grain price fluctuations to various climate change factors, as measured by correlation coefficient and R² of optimal regression equation, increased significantly in 1770s-1830s and 1870s-1890s. The influence of precipitation and drought on the fluctuation of grain price was more significant in these cold periods，due to the adverse effect of temperature on grain production. The per capita cultivated area continued to decrease before 1880, on the verge of threshold for subsistence level by 1780s, while the correlation coefficient between drought intensity and coverage indexes and detrended grain price, and R² of optimal regression equation reached an extremely significant level in most periods (except 1841-1865) (P<0.001). , When the grain reserves of the Qing government were greatly reduced and the financial balance was deteriorated in 1796-1815 and 1856-1880, the sensitivity of grain price to climate change was significantly enhanced, while the sensitivity was significantly reduced in 1736-1770, 1816-1855 and 181-1911, when the elasticity of economic subsystems was enhanced to a certain extent.

(4) The sensitivity of grain price to climate change in North China in the Qing Dynasty, was characterized by obvious regional difference and spatial aggregation, which was mainly affected by agricultural production conditions and water transport conditions. The correlation between the detrended grain price index and climate change factors at different spatial scales in North China, and R² of the optimal regression equations all showed similar spatial distribution pattern. These sensitivity indexes of grain price to climate change were higher in south Shanxi Province, south Zhili province, northwest and west Henan province, with decreasing trend to surrounding areas. There were significant differences (P<0.001) in the sensitivity indexes of grain price to climate change in different agricultural climate zones, such as Loess Plateau, Huang-Hai River Plain, lower Yellow River region, Huaibei-Ludong region. The regional differences of the sensitivity of grain price to local climate change factors were more significant. As an important factor related to grain market, water transport conditions can significantly affect the sensitivity of grain prices to climate change factors in the whole. The average correlation coefficient between differential grain price in the prefectures with water pathway and other regions, was significantly higher (P=0.004) than the prefectures with no water pathway. While mean correlation coefficients was significantly correlated with the sensitivity indexes of detrended grain price to climate change factors in the whole (r=0.726-0.860).

    There was an obvious correspondence between the phased changes, periodic fluctuation, spatial-temporal variation patterns of detrended grain price and famine in North China in the Qing Dynasty. There were also significant positive correlation and interaction between detrended grain price and famine. In addition to war and national-wide pandemic, the grain price volatility was an important factor affecting the epidemic in North China in Qing Dynasty, with obvious threshold for  its influence. In the periods when detrended grain prices were high or rising sharply, the famine counties indexes were often higher than the overall average. The detrended grain price index in North China had a periodic resonance with the famine counties indexes at time scale of 18-30 and 8-12 years, showing stable positive phase relationship, and higher significance in the cold period of 1781-1890. Most region-wide and local grain price anomalies had strong correspondence with the higher famine coverage indexes. The correlation coefficient between the famine indexes in North China and the detrended grain prices of the same year, and the lagged 2 years reached 0.420, 0.641 and 0.340 (P<0.001), the lagging effect of the famine index on detrended grain prices was still significant even after the auto-correlation of grain price was excluded. After the external imported epidemic disasters, and those related to wars in the years of 1820-1822, 1862-1864, 1888-1895, 1902, 181-1868 were excluded, the correlation coefficient between detrended grain price and epidemic index was 0.592 (P<0.001), the effect on the epidemic index was significantly enhanced (r=0.699) when detrended grain price exceeded certain limit.