Linking Remote Sensing and Dendrochronology to Quantify Climate‐Induced Shifts in High‐Elevation Forests Over Space and Time

Abstract

It is well known that tree growth is strongly affected by climate at high elevations, but it is still unclear how climate variability influences the distribution of montane forest ecosystems, particularly tree line species. To advance knowledge in this field, we combined temporal (tree ring measurements) and spatial data (remotely sensed variables), to quantify how forests responded to climatic variability across altitudinal gradients in Central Mexico. Normalized difference vegetation index (NDVI), tree‐ring chronologies, and site‐level climatic data were used in a vegetation trend analysis of greenness and browning and to reconstruct canopy vigor for the last century. Although a common ring width chronology was developed, we found significant site‐dependent forest growth response, where young trees (<100 years) exhibited heterogeneous growth trends, without an altitudinal pattern. However, mature trees (100–200 years) showed a common growth decline during the middle twentieth century, regardless of their altitude. Annual maximum NDVI anomalies did not show a general greening effect at high‐elevations. The forest showed both greening and browning zones denoting spatial variability in tree vigor. Furthermore, temperature from the previous year had a positive effect on both NDVI and ring width index but negative at beginning of the growing season. The significant relationship between winter‐spring NDVI (December to March) and ring width index (r = 0.64, p < 0.05), was useful to reconstruct the canopy vigor for the last 151 years. Thus, the greening effect derived from NDVI should be carefully interpreted as a direct forest growth increase. These results highlight the potential of integrating remotely sensed and dendrochronological methods to improve predictions of forest ecosystem responses.

From Trees to Ecosystems: Spatiotemporal Scaling of Climatic Impacts on Montane Landscapes Using Dendrochronological, Isotopic, and Remotely Sensed Data

“…This phenomenon was accentuated during summer 2011 when trees were under the most extreme VPD and temperature for the studied period. We also found that previous winter conditions (e.g., higher precipitation and warmer temperatures) can affect the isotopic signature of any given annual tree ring because the next growing season carries some of the photosynthetic materials from the previous year for early growth (Correa‐Díaz et al, 2019). Regarding spatial position, we found that δ 13 C ratios were less sensitive to previous climatic conditions at high elevations (south aspect) than middle elevation range or lower forest limit that agrees with Wu et al (2015) analysis (Table S3).…”

“…As a general overview across the entire chronology (1866–2016), low‐elevation sites (3500‐NW and 3500‐SW) have historically exhibited the highest average productivity, expressed as annual basal area increments (BAI), which until recently was up to four times higher than BAIs at high‐elevation sites (3900‐NW and 3900‐SW) (see Figure 8a and Correa‐Díaz et al (2019) for further information). However, changes in physiological performance (described above) led to a progressive downward trend in BAI since the mid‐20th century, which changed those historic site differences.…”

“…Our approach builds on previous studies that have found links between tree ring widths and remotely sensed data in México (Correa‐Díaz et al, 2019). We also took advantage of well‐established relationships identified in other parts of the world, such as the ones reported by Guo and Xie (2006) who found that August‐September NDVI has the same trend as δ 13 C of C 3 plant leaves in high‐altitudinal forests of the Tibetan Plateau.…”

“…For P. hartwegii , the growing season goes from the end of April to October (≈7 months) (Astudillo‐Sánchez et al, 2017; Biondi et al, 2005), but the most notable green‐up according to NDVI starts in May, which agrees with our results that show a maximum association with δ 13 C and Δ 13 C (Figures 7a and 7c). Indeed, Correa‐Díaz et al (2019) found that several climatic variables in May (minimum temperature, number of frost‐free days, and precipitation) were significantly associated with tree ring width. Besides that, the previous winter season was also important for δ 13 C and δ 18 O (Figures 7a and 7b); thus, higher productivity with warmer temperatures could generate photosynthetic reserves, which upon allocation to structural woody components undergo further isotope fractionation steps.…”

“…The quality control of dating and basal area increment (BAI) chronologies by site were done using the Dendrochronology Program Library in R (“ dplR ”) (Biondi & Qeadan, 2008; Bunn, 2008). To compare the forest growth trends with the physiological performance of P. hartwegii , we focused on the period 2000 to 2016 for BAI chronologies, but the complete time series extending back to the year 1664 can be found in Correa‐Díaz et al (2019).…”