No abstract

In the second paragraph of the section '2.3 Meteorological and remote sensing data', the Normalized Difference Vegetation Index (NDVI) data set used was incorrectly referred to as GIMMS-NDVI version 3G with a 0.084 • spatial resolution. This should be corrected to GIMMS-NDVI version G with a 0.07 • spatial resolution. N 2005 An extended AVHRR 8-km NDVI dataset compatible with MODIS and SPOT vegetation NDVI data Int. J.

Abstract. Ring-width (RW) records from the Gulf of Alaska (GOA) have yielded a valuable long-term perspective for North Pacific changes on decadal to longer timescales in prior studies but contain a broad winter to late summer seasonal climate response. Similar to the highly climatesensitive maximum latewood density (MXD) proxy, the blue intensity (BI) parameter has recently been shown to correlate well with year-to-year warm-season temperatures for a number of sites at northern latitudes. Since BI records are much less labour intensive and expensive to generate than MXD, such data hold great potential value for future treering studies in the GOA and other regions in mid-to high latitudes. Here we explore the potential for improving treering-based reconstructions using combinations of RW-and BI-related parameters (latewood BI and delta BI) from an experimental subset of samples at eight mountain hemlock (Tsuga mertensiana) sites along the GOA. This is the first study for the hemlock genus using BI data. We find that using either inverted latewood BI (LWB inv ) or delta BI (DB) can improve the amount of explained temperature variance by > 10 % compared to RW alone, although the optimal target season shrinks to June-September, which may have implications for studying ocean-atmosphere variability in the region. One challenge in building these BI records is that resin extraction did not remove colour differences between the heartwood and sapwood; thus, long term trend biases, expressed as relatively warm temperatures in the 18th century, were noted when using the LWB inv data. Using DB appeared to overcome these trend biases, resulting in a reconstruction expressing 18th-19th century temperatures ca. 0.5 • C cooler than the 20th-21st centuries. This cool period agrees well with previous dendroclimatic studies and the glacial advance record in the region. Continuing BI measurement in the GOA region must focus on sampling and measuring more trees per site (> 20) and compiling more sites to overcome site-specific factors affecting climate response and using subfossil material to extend the record. Although LWB inv captures the interannual climate signal more strongly than DB, DB appears to better capture long-term secular trends that agree with other proxy archives in the region. Great care is needed, however, when implementing different detrending options and more experimentation is necessary to assess the utility of DB for different conifer species around the Northern Hemisphere.

In the second paragraph of the section '2.3 Meteorological and remote sensing data', the Normalized Difference Vegetation Index (NDVI) data set used was incorrectly referred to as GIMMS-NDVI version 3G with a 0.084 • spatial resolution. This should be corrected to GIMMS-NDVI version G with a 0.07 • spatial resolution. N 2005 An extended AVHRR 8-km NDVI dataset compatible with MODIS and SPOT vegetation NDVI data Int. J.