Tracking phenological change in a regionally explicit context is a key to understanding ecosystem status and change. The current study investigated long-term trends of satellite-observed land surface phenology (LSP) in the 17 National Ecological Observatory Network (NEON) domains across the conterminous United States (CONUS). Characterization of LSP trends was based on a high temporal resolution (3-d) time series of the two-band enhanced vegetation index (EVI2) derived from a long-term data record (LTDR) of the Advanced Very High Resolution Radiometer (AVHRR) and the Moderate Resolution Imaging Spectroradiometer (MODIS). We identified significant trend patterns in LSP and their seasonal climate and land use/land cover drivers for each NEON domain. Key findings include (1) the start of season (SOS) predominantly shifted later in 13 out of 17 domains (24.3% of CONUS by area) due potentially to both a lack of spring warming in the eastern United States and changes in agronomic practices over agricultural lands; (2) the end of season (EOS) became predominantly later in nine domains dominated by natural vegetation (14.1% of CONUS by area) in response to widespread warming in autumn; (3) the EOS predominantly shifted earlier in three domains (10.6% of CONUS by area) over primarily agricultural lands as potentially affected by changes in crop growth cycles; and (4) earlier shift in the SOS was mostly found in the Northwest (3.6% of CONUS by area) and was predominant only in the moist Pacific Northwest (27.7% of the domain by area) in response to more pronounced spring warming in the region. The overall patterns of SOS and EOS trends across CONUS appeared constrained by continental-scale temperature trends as characterized by a west-east dipole and the distribution of the nation's agricultural lands. In addition, seasonal trend analysis revealed that most NEON domains (15/17) became predominantly greener in part of or throughout the growing season, potentially contributed by both climate change-induced growth increase and improved agricultural productivity. The domain-wide LSP trends with their underlying drivers identified here provide important contextual information for NEON science as well as for investigations within CONUS using other distributed observatories (e.g., LTER, LTAR, FLUXNET, USA-NPN, etc.).

中文翻译：

---

通过 NEON 域分析的美国本土 (1982-2016) 地表物候趋势

在区域明确的背景下跟踪物候变化是了解生态系统状态和变化的关键。目前的研究调查了美国本土 (CONUS) 17 个国家生态观测网络 (NEON) 域中卫星观测地表物候 (LSP) 的长期趋势。LSP 趋势的表征基于来自高级甚高分辨率辐射计 (AVHRR) 的长期数据记录 (LTDR) 的双波段增强植被指数 (EVI2) 的高时间分辨率 (3-d) 时间序列和中分辨率成像光谱仪 (MODIS)。我们为每个 NEON 域确定了 LSP 及其季节性气候和土地利用/土地覆盖驱动因素的重要趋势模式。主要发现包括 (1) 季节开始 (SOS) 在 17 个领域中的 13 个领域（按地区划分的美国本土的 24.3%）主要转移到较晚，这可能是由于美国东部缺乏春季变暖以及过去几年农艺实践的变化。农业用地；(2) 在秋季普遍变暖的情况下，季节结束 (EOS) 在以自然植被为主的 9 个区域（按面积占 CONUS 的 14.1%）主要变得较晚；(3) EOS 主要在三个领域（按面积占美国本土的 10.6%）提前转移，主要是农业用地，这可能会受到作物生长周期变化的影响；(4) SOS 较早的变化主要出现在西北地区（按面积占 CONUS 的 3.6%），仅在潮湿的太平洋西北部（按面积占面积的 27.7%）占主导地位，以响应更明显的春季变暖。地区。整个 CONUS 的 SOS 和 EOS 趋势的整体模式似乎受到大陆尺度温度趋势的限制，其特征是东西向偶极子和国家农业用地的分布。此外，季节性趋势分析显示，大多数 NEON 域 (15/17) 在部分或整个生长季节变得更加绿色，这可能是由气候变化引起的增长增加和农业生产力提高造成的。域范围的 LSP 趋势及其在此确定的潜在驱动因素为 NEON 科学以及 CONUS 内使用其他分布式天文台（例如，LTER、LTAR、FLUXNET、USA-NPN 等）的调查提供了重要的背景信息。