Brain Korea 21 Plus Team, Seoul National University, Seoul, Korea.
Department of Landscape Architecture and Rural Systems Engineering, Seoul National University, Seoul, Korea.
Glob Chang Biol. 2017 Oct;23(10):4133-4146. doi: 10.1111/gcb.13787. Epub 2017 Jul 6.
Understanding the long-term performance of global satellite leaf area index (LAI) products is important for global change research. However, few effort has been devoted to evaluating the long-term time-series consistencies of LAI products. This study compared four long-term LAI products (GLASS, GLOBMAP, LAI3g, and TCDR) in terms of trends, interannual variabilities, and uncertainty variations from 1982 through 2011. This study also used four ancillary LAI products (GEOV1, MERIS, MODIS C5, and MODIS C6) from 2003 through 2011 to help clarify the performances of the four long-term LAI products. In general, there were marked discrepancies between the four long-term LAI products. During the pre-MODIS period (1982-1999), both linear trends and interannual variabilities of global mean LAI followed the order GLASS>LAI3g>TCDR>GLOBMAP. The GLASS linear trend and interannual variability were almost 4.5 times those of GLOBMAP. During the overlap period (2003-2011), GLASS and GLOBMAP exhibited a decreasing trend, TCDR no trend, and LAI3g an increasing trend. GEOV1, MERIS, and MODIS C6 also exhibited an increasing trend, but to a much smaller extent than that from LAI3g. During both periods, the R of detrended anomalies between the four long-term LAI products was smaller than 0.4 for most regions. Interannual variabilities of the four long-term LAI products were considerably different over the two periods, and the differences followed the order GLASS>LAI3g>TCDR>GLOBMAP. Uncertainty variations quantified by a collocation error model followed the same order. Our results indicate that the four long-term LAI products were neither intraconsistent over time nor interconsistent with each other. These inconsistencies may be due to NOAA satellite orbit changes and MODIS sensor degradation. Caution should be used in the interpretation of global changes derived from the four long-term LAI products.
了解全球卫星叶面积指数（LAI）产品的长期性能对于全球变化研究非常重要。然而，很少有研究致力于评估 LAI 产品的长期时间序列一致性。本研究比较了四个长期 LAI 产品（GLASS、GLOBMAP、LAI3g 和 TCDR）在 1982 年至 2011 年期间的趋势、年际变异性和不确定性变化。本研究还使用了四个辅助 LAI 产品（GEOV1、MERIS、MODIS C5 和 MODIS C6）在 2003 年至 2011 年期间，以帮助澄清四个长期 LAI 产品的性能。总体而言，四个长期 LAI 产品之间存在明显差异。在 MODIS 之前的时期（1982-1999 年），全球平均 LAI 的线性趋势和年际变异性遵循 GLASS>LAI3g>TCDR>GLOBMAP 的顺序。GLASS 的线性趋势和年际变异性几乎是 GLOBMAP 的 4.5 倍。在重叠期间（2003-2011 年），GLASS 和 GLOBMAP 呈下降趋势，TCDR 无趋势，LAI3g 呈上升趋势。GEOV1、MERIS 和 MODIS C6 也呈上升趋势，但幅度要小得多。在两个时期，四个长期 LAI 产品之间去趋势异常的 R 值在大多数地区都小于 0.4。两个时期的四个长期 LAI 产品的年际变异性都有很大差异，其差异顺序为 GLASS>LAI3g>TCDR>GLOBMAP。通过配置误差模型量化的不确定性变化也遵循相同的顺序。我们的结果表明，四个长期 LAI 产品既不是时间内一致的，也不是相互一致的。这些不一致可能是由于 NOAA 卫星轨道变化和 MODIS 传感器退化所致。在解释来自四个长期 LAI 产品的全球变化时应谨慎行事。
