为探究全球气候变化条件下松花江流域陆地生态系统健康程度的变化特征，基于2000—2020年MODIS MOD17A3HGF数据集，采用趋势分析、相关性分析、M-K检验、地理探测器和相对重要性分析等方法，结合气象站点数据和土地利用数据，分析植被净初级生产力（net primary productivity，NPP）时空演变特征及其对极端气候事件的响应机制。结果表明：松花江流域年均NPP值为407.45g/m2（以C计，下同），以年均4.82g/m2的速率显著上升（p<0.01）；极端降水事件对植被NPP空间分异性的影响强于极端气温事件，极端气候指数间交互作用的影响大于单一极端气候指数的影响，流域及农田和草地生态系统NPP主要受总降水量（PRCPTOT）与年平均最低气温（TMIN）交互作用的影响，森林、湿地和聚落生态系统NPP分别受中雨日数（R10mm）与年平均最高气温（TMAX）交互作用、强降水量（R95P）与TMIN交互作用和 R10mm与暖夜日数（TN90P）交互作用的影响；时间尺度上PRCPTOT、TMAX和TMIN是植被NPP的主要影响因素，空间尺度上PRCPTOT和TMIN是多年平均NPP的主要影响因素。研究结果可为量化气候变化背景下区域生态系统健康程度和应对极端气候事件措施的制定提供科学依据。

The carbon sequestration capacity of vegetation over 21 years (2000-2020) was analyzed using MODIS MOD17A3HGF datasets and the spatio-temporal evolution characteristics of net primary productivity (NPP) was investigated to explore the changing characteristics of terrestrial ecosystem health in the Songhua River basin under the condition of global climate change. The reaction of NPP to anomalous climate incidents was analyzed through using data on daily precipitation, maximum temperature, and minimum temperature from 80 regular meteorological stations situated in the Songhua River basin and its adjacent regions. The results could provide a scientific basis for quantifying the health of regional ecosystems in the context of climate change and for the development of measures to cope with extreme climate events. A variety of research methods were used such as correlation analysis, Mann-Kendall (M-K) test, GeoDetector, and relative importance ranking. The trend analysis techniques adopted encompassed one-way linear regression and Theil-Sen Median trend analysis. One-way linear regression was used to examine the linear trend of annual mean NPP within the Songhua River basin, while Theil-Sen Median trend analysis was used to examine the dynamic evolution of spatial patterns of NPP within the Songhua River basin. Pearson correlation coefficients was calculated based on pixel values to assess the relationship between vegetation NPP and extreme climate indices. The M-K test was utilized to establish the statistical significance of NPP trends, with the resulting outcome conveyed through the Z statistic. The GeoDetector parameter was employed, consisting of a "Factor detector", "Interaction detector", and "Risk detector", to explore the impact of extreme weather indices on vegetation NPP across the Songhua River basin. Multiple regression methods were utilized to investigate the effect of extreme climate indices on vegetation NPP in the Songhua River basin, considering both temporal and spatial scales. The results show that the average annual NPP within Songhua River basin showed an oscillating upward trend from 2000 to 2020, annually increasing at a rate of 4.82 g/m2 (calculated by C, same below). The annual NPP varied from 315.48 to 464.38 g/m2, annually averaging at 407.45 g/m2 over the 21 years. The maximum value occurred in 2014, reaching a peak of 464.38 g/m2, and the minimum value was observed in 2000 at 315.48 g/m2. Forest ecosystems had the highest mean annual NPP value, standing at 521.73 g/m2. Grassland and agroecosystems followed with the second-highest mean annual NPP values of 378.38 g/m2 and 343.26 g/m2, respectively. On the other hand, colony and wetland ecosystems had relatively lower mean annual NPP values, reaching 331.26 and 308.75 g/m2, respectively. The grassland ecosystem showed the most rapid growth rate, annually increasing by 5.64 g/m2 , closely followed by forest ecosystems, which exhibited an increase rate of 5.61 g/m2. In contrast, wetland ecosystems displayed the slowest increase rate at 3.44 g/m2. Regarding spatial distribution, the vegetation NPP within Songhua River basin showed an irregular pattern with high values in the southeast, low values in the southwest, high values in the surrounding areas, and low values in the central region. The areas with high annual mean NPP were mainly concentrated in Harbin City and Mudanjiang City in Heilongjiang Province and Jilin City, Yanbian Korean Autonomous Prefecture, and Baishan City in Jilin Province. In these areas, most of the flora displayed annual NPP values above 500 g/m2, with annual mean NPP levels also surpassing 500 g/m2. The strong positive correlation between the vegetation NPP and the extreme precipitation indices in the Songhua River basin was more significant than the corresponding negative correlations. Importantly, NPP illustrated predominant and statistically meaningful correlations with PRCPTOT,R10mm, and R95P. The significance of vegetation NPP in relation to extreme temperature events within Songhua River basin was relatively low, indicating that the extreme temperature indices did not significantly limit the ability of vegetation to sequester carbon compared to the extreme precipitation indices.