Spectroscopic analysis of seasonal changes in live fuel moisture content and leaf dry mass

REMOTE SENSING OF ENVIRONMENT

卷:150

Article

Qi, Yi¹  Dennison, Philip E.¹  Jolly, W. Matt²  Kropp, Rachael C.²  Brewer, Simon C.¹ 

[1] Univ Utah, Dept Geog, Salt Lake City, UT 84112 USA

[2] Missoula Fire Sci Lab, Rocky Mt Res Stn, Missoula, MT 59808 USA

关键词: Live fuel moisture content;    Leaf dry mass;    Spectroscopy;    Partial least squares regression;   

DOI  :  10.1016/j.rse.2014.05.004

来源: Elsevier

【 摘 要 】

Live fuel moisture content (LFMC), the ratio of water mass to dry mass contained in live plant material, is an important fuel property for determining fire danger and for modeling fire behavior. Remote sensing estimation of LFMC often relies on an assumption of changing water and stable dry mass over time. Fundamental understanding of seasonal variation in plant water and dry mass is needed to explain the spectral expression of LFMC changes over time. We conducted a five-month experiment to continuously measure field LFMC samples, biochemical components of dry matter, and leaf spectroscopic data for two species common in the western U.S., lodgepole pine (Pinus contorta Douglas ex Loudon) and big sagebrush (Artemisia tridentate Nutt). Our results showed that new lodgepole pine needles initially had higher LFMC and a smaller proportion of dry mass, but differences between new and old needles converged as the new needles matured. New needle dry mass had strong temporal trends, and dry mass explained more variation in LFMC than water in both new and old needles. Sagebrush leaves exhibited decreasing trends in LFMC, but water and dry mass comparably contributed to LFMC seasonal variation. Spectroscopic analysis using partial least squares regression (PLSR) showed good modeling accuracy for LFMC temporal variation in new needles (R-2 = 0.94, RMSE = 5.84%), old needles (R-2 = 0.72, RMSE = 3.51%), and sagebrush (R-2 = 0.91, RMSE = 21.03%). Spectral variation in response to changing LFMC and dry mass was difficult to isolate from broader spectral trends due to chlorophyll absorption, leaf structure, water absorption, and co-varied biochemical components. Our results stress cautious spectral interpretation and wavelength selection for LFMC estimation in some species (e.g. lodgepole pine), since temporal changes in spectra may dominantly reflect temporal variation in dry mass, pigments, and/or structure rather than water content. Since new needles should have stronger spectral expression at the canopy scale, differing temporal trends in new and old lodgepole pine needles provides an additional complicating factor for remote monitoring of LFMC. (C) 2014 Elsevier Inc All rights reserved.

【 授权许可】

Free