Abstract

Vertically integrated atmospheric water vapor (VIWV) over the Indo-Pacific warm pool (IPWP) indirectly affects terrestrial vegetation growth (TVG) patterns through atmospheric water vapor transmission. However, their linkages and mechanisms are poorly understood. This study intends to understand the contributions of VIWVIPWP to TVG and the mechanisms by which VIWVIPWP impacts TVG. Combining monthly SST, VIWV, and NDVI data from 1982 to 2015, the linkage between VIWVIPWP and NDVI is investigated during April–June (AMJ). A strong correlation between VIWVIPWP and NDVI suggests that VIWVIPWP is an important factor affecting TVG. A composite analysis of VIWVIPWP anomalies and their relation to NDVI patterns shows that VIWVIPWP positively influences the NDVI of 68.1% of global green land during high-VIWVIPWP years but negatively influences 74.7% in low years. Corresponding to these results, during high-VIWVIPWP years, the warm and humid terrestrial climate conditions improved TVG by 9% and 2% in the Northern and Southern Hemispheres, respectively, but cold and dry conditions inhibited TVG for both hemispheres during the low years. Additionally, strong spatial correlations between VIWVIPWP and precipitation imply that VIWVIPWP affects the spatial–temporal pattern of precipitation. There is a stronger interaction between the Pacific north–south ridge and the two land troughs during high-VIWVIPWP years than during low-VIWVIPWP years. The zonally averaged wind at 850 hPa and VIWV results indicate that, during high-VIWVIPWP years, the enhanced wind from the ocean brings more atmospheric water vapor to land, increasing the probability of precipitation and resulting in moist climate conditions that promote AMJ vegetation growth. In brief, VIWVIPWP indirectly induces vegetation growth by affecting the distributions of terrestrial VIWV and precipitation.

1. Introduction

As a globally unique place, the Indo-Pacific warm pool (IPWP) over the Indian Ocean and the Pacific Ocean is a key site for the global hydrologic cycle. It is the largest source of atmospheric water vapor, the largest warm pool on Earth (Niedermeyer et al. 2014), and is called the steam engine of the world (De Deckker 2016). The IPWP is characterized by a permanent sea surface temperature (SST) of ≥28.5°C (Huang and Mehta 2004) in the global ocean; its location is depicted in Fig. 1. Here, intense atmospheric convection and rainfall provide a major driving force for global atmospheric circulation (Niedermeyer et al. 2014). Atmospheric convection over the IPWP can influence the global distribution of heat and water vapor by generating planetary waves (Zhou et al. 2018). For instance, in the tropics (Feng and Li 2013; Ma and Li 2008) and at higher latitudes, previous studies have found that the IPWP significantly affects climatic conditions. In East Asia, IPWP effects have been studied extensively, with results showing that the monsoons (Annamalai et al. 2013; D’Arrigo et al. 2006; Feng et al. 2011; Shan et al. 2014; Huang and Sun 1992; Shin et al. 2011; Li et al. 2013) are all closely linked to IPWP variability. In East Africa, spring droughts are also related to IPWP SST variability (Funk et al. 2014). Additionally, the IPWP affects global-scale climate change (Visser et al. 2003; Zhou et al. 2018). These studies show that the IPWP has a strong influence on global heat and water vapor transport and thus constitutes an important component of the climate system (Russell et al. 2014; Visser et al. 2003; Zhu et al. 2012).

Fig. 1.

Map of the IPWP area. The IPWP area of interest is indicated by the shaded area. The IPWP is characterized by the annual-averaged sea surface temperature (SST) ≥ 28.5°C (Huang and Mehta 2004) in the global ocean over 1982–2015.

Fig. 1.

Map of the IPWP area. The IPWP area of interest is indicated by the shaded area. The IPWP is characterized by the annual-averaged sea surface temperature (SST) ≥ 28.5°C (Huang and Mehta 2004) in the global ocean over 1982–2015.

Changes in the IPWP can affect the Walker and Hadley circulations by locally altering the divergent atmospheric flow, causing a global climate response (Neale and Slingo 2003; Sardeshmukh and Hoskins 1988). Therefore, the IPWP is closely related to large-scale atmospheric circulations (LACs), which are important driving forces of atmospheric water vapor transmission. LACs induce the movement of atmospheric water vapor originating from the ocean to the land and further affect the distribution of global atmospheric water vapor. LACs also have major influences on local precipitation variations. High water vapor over land increases the probability of precipitation and further affects the geographical distribution of terrestrial precipitation. Precipitation is the end product of water vapor movement and changes over years. For example, atmospheric water vapor originating from the Indo-Pacific region dominates the precipitation patterns in southeastern Asia, while water vapor from the western Pacific partially contributes to the precipitation in East Asia. Obviously, atmospheric water vapor changes represent an important tool in the study of climate, the hydrological cycle, and vegetation growth. Thus, changes in the vertically integrated atmospheric water vapor over the IPWP (VIWVIPWP) can be used as an important signal for projecting global distribution patterns of moisture and vegetation.

Importantly, global vegetation growth is sensitive to the variability of climate variables (e.g., precipitation). Precipitation is one of the key hydrometeorological factors affecting the spatial–temporal pattern of terrestrial vegetation growth (TVG). Precipitation availability is a rather ubiquitous control of the spatial distribution of vegetation growth, so its variations cause significant changes in the percent coverage of green land. Furthermore, critical water deficits cause the soil water potential to drop, causing cavitation and plant mortality. Changes in precipitation have more complex effects on vegetation. As previously mentioned, precipitation covaries with atmospheric water changes via ocean–atmosphere–land interactions, thus resulting in changes in vegetation growth. In brief, VIWVIPWP has a larger direct impact on the spatiotemporal distribution of global atmospheric water vapor and precipitation, and therefore indirectly affects TVG patterns through precipitation. However, the mechanism of indirect influence of ocean water vapor on TVG through atmospheric transmission is poorly understood. Notably, relatively little attention has been paid to how atmospheric water vapor indirectly induces vegetation growth by regulating the initial conditions of precipitation and thus affecting its patterns. It is therefore important to study how VIWVIPWP changes influence global vegetation growth.

Previous studies have mainly focused on the influence of the IPWP on weather and climate over regional or small scales, and less attention has been paid to the effects of the IPWP on global vegetation growth. On a global scale, we first investigate the effects of the IPWP on climate and TVG, which is the best expression of the IPWP function as a global “heat and moisture engine.” Based on the mechanisms of the response of vegetation growth to precipitation activities, we detected characteristics of the global vegetation growth response to temporal and spatial changes in precipitation regulated by IPWP-induced anomalies in global atmospheric water vapor changes. This study will provide a useful knowledge framework for understanding the long-range impact of the IPWP on the function of terrestrial ecosystems via the water vapor engine.

Here, we use a large set of VIWV and global NDVI (a proxy of vegetation growth activity) data to ascertain the relationship between vegetation growth activity and VIWVIPWP changes during typical seasons from 1982 to 2015. In addition, we investigate anomalous VIWVIPWP modes in conjunction with modes of spatiotemporal variation in vegetation growth activities, temperature, and precipitation. We also investigate the linkages between atmospheric circulation/meteorological factors (sea level pressure, wind field at 850 hPa, and column-atmospheric water vapor) and vegetation activity. Finally, we reveal the probable physical mechanisms for the linkages. As the physical mechanism of the remote linkage between ocean VIWV and TVG remains unclear, the purpose of this study is to test the hypothesis that the VIWV anomaly over the key area of the global ocean affects the distribution of global VIWV through LACs, thus regulating the distribution of terrestrial precipitation and indirectly inducing terrestrial vegetation growth at key stages of rapid growth during the growth season. These results will help deepen the understanding of the probable physical mechanisms of ocean–atmosphere–vegetation interactions through atmospheric water vapor transmission–induced terrestrial precipitation.

2. Data and methods

a. Vegetation data

Global vegetation activity patterns were investigated by using normalized difference vegetation index (NDVI) datasets. These datasets were extracted from GIMMS NDVI3g.v1 (third-generation, version 1) biweekly products with a spatial resolution of 8 km × 8 km. The datasets from July 1981 to December 2015 were downloaded (from https://ecocast.arc.nasa.gov/data/pub/gimms/3g.v1/00FILE-LIST.txt). The NDVI3g.v1 dataset was calibrated and corrected for view geometry, volcanic aerosols, and other effects unrelated to vegetation change. The NDVI is calculated from the following formula:

 
formula

where NIR and VIS indicate the spectral reflectance measurements acquired in the near-infrared (NIR) and visible (VIS) regions, respectively (see https://earthobservatory.nasa.gov/features/MeasuringVegetation/measuring_vegetation_2.php). The NDVI datasets contain global geographical projections (U.S. Defense Mapping Agency 1987). GIMMS NDVI3g.v1 continuous data over the period from 1982 to 2015 were reconstructed by using maximum value compositing (MVC) (Holben 1986). The NDVI data vary from −1.0 to +1.0, where areas with high vegetation coverage and activity values are denoted by positive values, and ice and snow fields usually result in negative values; however, in desert regions, ground conditions (e.g., soil; Huete 1988) strongly affect the NDVI, leading to unstable values that do not accurately represent the vegetation status (Fang et al. 2004). Therefore, for the present analysis, we selected pixels with an NDVI value of more than 0.1 to represent vegetation density, and pixels with a mean NDVI < 0.1 were excluded (Zhou et al. 2003, 2001). In this way, we improved the credibility of the NDVI data and clarified the correlation between global vegetation and VIWVIPWP.

b. Climate data

The European Centre for Medium-Range Weather Forecasts interim reanalysis dataset (Dee et al. 2011; Simmons 2006) for the period from 1982 to 2015 (the same time span as the NDVI data) was employed in this study; the utilized data included monthly mean sea surface level pressure (SLP), VIWV, and the uυ wind vector at 850 hPa. The monthly mean global SST data from the COBE-SST2 dataset (Hirahara et al. 2014) with a 1° × °1 resolution were also used. The long-term monthly mean surface temperature and precipitation data available from the Climate Research Unit, version 3.24 (CRU 3.24), dataset (Jones and Harris 2013) with a 0.5° × 0.5° resolution were used for the period from 1982 to 2015.

c. Definition of VIWVIPWP events

We identified all the high-, normal-, and low-water-vapor events over the IPWP. Years with a VIWVIPWP value greater than (or equal to) 0.95 standard deviation) (SD) or less than (or equal to) −0.95 SD from the mean were defined as high- and low-VIWVIPWP years, respectively. According to this criterion, the high-VIWVIPWP years (1990, 1991, 1998, 2001, 2007, and 2011) and low-VIWVIPWP years (1982, 1983, 1984, 1985, 1987, and 1992) were determined. Additionally, based on the criterion of a VIWVIPWP value greater than −0.95 SD and less than 0.95 SD, the normal VIWVIPWP years (1986, 1988, 1989, 1993, 1994, 1995, 1996, 1997, 1999, 2000, 2002, 2003, 2004, 2005, 2006, 2008, 2009, 2010, 2012, 2013, 2014, and 2015) were identified.

d. Statistical analysis methods

To investigate the links between VIWVIPWP and NDVI, we conducted correlation analyses. The statistical significance is determined by the Student’s t test. We also used composite analysis, a method widely used in atmospheric sciences research as an important data analysis method to highlight the impact of VIWVIPWP on global vegetation. In addition, we defined an impact index [I-index; see (2) and (3)] to evaluate the impact of VIWVIPWP on the global NDVI. The formulas below were applied to the above evaluation:

 
formula
 
formula

where h, n, and l indicate high, normal, and low years, respectively; I-indexh and I-indexl are percentage change anomalies, and Vh, Vl, and Vn indicate anomalies in variables such as VIWVIPWP and the NDVI. Meteorological elements associated with high- and low-VIWVIPWP years were addressed by a composite analysis.

3. Results

a. The annual cycles of IPWP and the global NDVI

To accurately investigate the linkage between VIWVIPWP and global vegetation growth, we first detected the annual-change features of the IPWP and global vegetation growth with an analysis of SSTIPWP, VIWVIPWP, and NDVI (as a proxy for vegetation growth) features (see Fig. 2). Figure 2 shows that during January–December, SSTIPWP (see Fig. 2a) and VIWVIPWP (see Fig. 2b) were obviously higher than the 12-month average value. The maximum values of SSTIPWP and VIWVIPWP appeared in May and July, respectively. As shown in Fig. 2c, vegetation growth activities can be divided into a growing season (April–October) and a nongrowing season (November–March), with the minimum NDVI value occurring in February and the maximum value in July. Notably, there was a pronounced increase of 0.13 in NDVI from April to July, which accounts for 81.3% of the total increase seen in the NDVI (0.16 between February and July). Figure 3 also shows that there is vigorous vegetation growth activity during these months. This suggests that the vegetation activity during April–June (AMJ) is the most vigorous.

Fig. 2.

Annual (a) SST and (b) VIWV cycles over the IPWP region and (c) the global NDVI. The reference period is from 1982 to 2015. Yellow shading indicates the main research season (i.e., AMJ) defined in this paper.

Fig. 2.

Annual (a) SST and (b) VIWV cycles over the IPWP region and (c) the global NDVI. The reference period is from 1982 to 2015. Yellow shading indicates the main research season (i.e., AMJ) defined in this paper.

Fig. 3.

Differences between the monthly NDVI (δNDVI) of the current and previous month during April–July from 1982 to 2015. The δNDVI in April is the difference in the monthly NDVI between April and March. The dashed gray line indicates the average value for April–July.

Fig. 3.

Differences between the monthly NDVI (δNDVI) of the current and previous month during April–July from 1982 to 2015. The δNDVI in April is the difference in the monthly NDVI between April and March. The dashed gray line indicates the average value for April–July.

Additionally, AMJ is the season with the highest SSTIPWP and high VIWVIPWP. Furthermore, during AMJ the positive correlation between VIWVIPWP and SSTIPWP is at the 0.01 significance level (see Fig. 4), which implies that AMJ VIWVIPWP accurately reflects the importance of the IPWP as the key steam engine of the world. We also calculated the autocorrelations of VIWVIPWP and NDVI in this season (Table 1). The correlation between April and June VIWVIPWP is at the 90% significance level, while other values of VIWVIPWP and NDVI are all at the 99% significance level, as estimated from a standard Student’s t test.

Fig. 4.

Month-to-month relationship between VIWVIPWP and global NDVI during AMJ from 1982 to 2015.

Fig. 4.

Month-to-month relationship between VIWVIPWP and global NDVI during AMJ from 1982 to 2015.

Table 1.

The autocorrelations of VIWVIPWP and NDVI for April–June during 1982–2015. One asterisk signifies correlation at the 90% level, estimated by a local Student’s t test; two asterisks signify the 99% level.

The autocorrelations of VIWVIPWP and NDVI for April–June during 1982–2015. One asterisk signifies correlation at the 90% level, estimated by a local Student’s t test; two asterisks signify the 99% level.
The autocorrelations of VIWVIPWP and NDVI for April–June during 1982–2015. One asterisk signifies correlation at the 90% level, estimated by a local Student’s t test; two asterisks signify the 99% level.

According to the following information, the AMJ period strongly reflects key features of the IPWP and contains global vegetation growth activity. Importantly, the strongest activity response to the highest VIWVIPWP can be easily highlighted during AMJ, which would test the hypothesis that VIWVIPWP affects global vegetation growth activity. Therefore, we chose AMJ from 1982 to 2015 as the research period in this paper.

b. The linkage between VIWVIPWP and global NDVI during AMJ

To detect the linkage between VIWVIPWP and global NDVI, we made a simple correlation between AMJ VIWVIPWP and NDVI from 1982 to 2015 (Table 2). Comparing the correlations between AMJ VIWVIPWP and globally averaged NDVI for all seasons (Table 2), a higher positive correlation at the 99% significance level occurs during AMJ. This means that the global NDVI was more strongly and closely associated with VIWVIPWP during AMJ than during other seasons [except October–December (OND)]. During OND, the highest correlation between AMJ VIWVIPWP and the NDVI has no biophysical meaning of merit because most of the Northern Hemisphere (NH) is in a nongrowing season. We further investigated the relationship between intermonthly VIWVIPWP and NDVI during AMJ from 1982 to 2015 (Fig. 5). A strong and pronounced positive correlation (R = 0.603, p < 0.001) is shown in Fig. 5, which indicates that the global NDVI was more strongly and closely associated with VIWVIPWP during AMJ.

Table 2.

The correlations between NDVI and AMJ VIWVIPWP from 1982 to 2015. One asterisk signifies correlation at the 90% level, estimated by a local Student’s t test; two asterisks signify the 99% level.

The correlations between NDVI and AMJ VIWVIPWP from 1982 to 2015. One asterisk signifies correlation at the 90% level, estimated by a local Student’s t test; two asterisks signify the 99% level.
The correlations between NDVI and AMJ VIWVIPWP from 1982 to 2015. One asterisk signifies correlation at the 90% level, estimated by a local Student’s t test; two asterisks signify the 99% level.
Fig. 5.

The relationship between intermonthly VIWVIPWP and global NDVI during AMJ from 1982 to 2015. In axes’ titles, ∆ indicates a difference between the values of the current month and the previous month.

Fig. 5.

The relationship between intermonthly VIWVIPWP and global NDVI during AMJ from 1982 to 2015. In axes’ titles, ∆ indicates a difference between the values of the current month and the previous month.

The month-to-month changes in VIWVIPWP and the globally averaged NDVI during AMJ from 1982 to 2015 (see Fig. 6) show that the variation in VIWVIPWP is generally aligned with that of the globally averaged NDVI. The correlation coefficients between the series are 0.62 (containing the linear trend; not shown) and 0.69 (detrended); both are significant at the 99% statistical confidence level estimated by a local Student’s t test. This implies that the AMJ VIWVIPWP is an important factor that influences AMJ vegetation growth activity on a global scale, as it can be used to explain some of the AMJ NDVI variance. In addition, we also computed spatial correlations between VIWVIPWP and global NDVI (Fig. 7). As shown in Fig. 7, strong and pronounced positive correlations covered most of the NH, but strong and pronounced negative correlations were present in southern Africa, south-central South America, and northern Australia in the Southern Hemisphere (SH). Notably, on a global scale during AMJ, VIWVIPWP has a considerable diverse regional influence on vegetation growth. Therefore, we focused on the effect of VIWVIPWP on global vegetation growth during AMJ.

Fig. 6.

Changes in detrended NDVI and VIWVIPWP during AMJ from 1982 to 2015.

Fig. 6.

Changes in detrended NDVI and VIWVIPWP during AMJ from 1982 to 2015.

Fig. 7.

Spatial correlation coefficient between VIWVIPWP and the NDVI during AMJ from 1982 to 2015. Coefficients greater than +0.2 or less than −0.2 are significant at the 95% confidence level.

Fig. 7.

Spatial correlation coefficient between VIWVIPWP and the NDVI during AMJ from 1982 to 2015. Coefficients greater than +0.2 or less than −0.2 are significant at the 95% confidence level.

c. The effect of VIWVIPWP anomalies on the global NDVI

To accurately evaluate the effect of VIWVIPWP on global vegetation growth, we focused on anomalous VIWVIPWP years to highlight these effects. Therefore, a composite analysis was conducted to identify the AMJ NDVI anomaly patterns through the investigation of NDVI differences between the high (low)-VIWVIWIP and normal years (Fig. 8). During high-VIWVIPWP years (Fig. 8a), the positive NDVI data cover up to 68.1% of global green land. Their corresponding regions include the following: most of Europe, southern Africa, central and eastern Australia, northeast North America, and most of South America. The negative NDVI data during low-VIWVIPWP years (Fig. 8b) compared with those in normal years covered 74.7% of global vegetation-covered areas. Their corresponding regions are larger than those in high years. In addition, we investigated these effects on the global NDVI during anomalous years (Fig. 9) using the I-index. The impact of VIWVIPWP on global NDVI anomalies varied across the VIWVIPWP years. For example, during high-VIWVIPWP years, VIWVIPWP had an evident positive impact on the NDVI of up to 68.1% of vegetation-covered areas worldwide; in these years, the NDVI tended to be higher than that in normal years and was accompanied by an increase of 9% and 2% in the NDVI in the NH and SH, respectively. In contrast, VIWVIPWP had a negative impact on the NDVI of up to 74.7% of vegetation-covered areas worldwide in years when the NDVI tended to be lower than that in normal years and was accompanied by a decrease of 4% and 3% in the NDVI in the NH and SH, respectively. Notably, the geographical pattern of the impact of high-VIWVIPWP years on the NDVI was different than that for low-VIWVIPWP years.

Fig. 8.

(a) AMJ differences in NDVI between the high and normal years. (b) As in (a), but for the low years. The difference values represent the difference between the high- or low-VIWVIPWP years and the normal-VIWVIPWP years.

Fig. 8.

(a) AMJ differences in NDVI between the high and normal years. (b) As in (a), but for the low years. The difference values represent the difference between the high- or low-VIWVIPWP years and the normal-VIWVIPWP years.

Fig. 9.

Effect of VIWVIPWP anomalies on the global NDVI during (a) high and (b) low years. These percentages were calculated based on the I-index.

Fig. 9.

Effect of VIWVIPWP anomalies on the global NDVI during (a) high and (b) low years. These percentages were calculated based on the I-index.

d. The climate factor features associated with high and low-VIWVIPWP years

To investigate the causes underlying the impact of VIWVIPWP on the global NDVI, we chose two main climate factors (temperature and precipitation) closely related with vegetation growth and mapped their spatial distribution features during high- and low-VIWVIPWP years. Figure 10 depicts the spatial changes in temperature and precipitation anomalies during high- and low-VIWVIPWP years. We found that during high-VIWVIPWP years, there was an increase in temperature over most of the global area (Fig. 10a), but for low-VIWVIPWP years the opposite was true (Fig. 10b). In addition, we found that an increase in precipitation occurred for most parts of Eurasia and South America, central and eastern North America, central South Africa, and Australia during high-VIWVIPWP years (Fig. 10c), but for low-VIWVIPWP years (Fig. 10d) the precipitation was lower in the same regions. Notably, there was an apparent overall spatial increase (or overall spatial decrease) in global temperature during high (or low) VIWVIPWP years, but precipitation showed strong diverse spatial features of increase (or decrease). As mentioned previously, warmer, wetter climate conditions during high-VIWVIPWP years affected most of the land. In contrast, during low-VIWVIPWP years, cooler, drier climate conditions appeared over the regions listed previously.

Fig. 10.

Composite temperature anomalies for (a) high- and (b) low-VIWVIPWP years; (c),(d) as in (a) and (b), respectively, but for precipitation. The anomaly values represent the difference between the high- or low-VIWVIPWP years and the normal-VIWVIPWP years.

Fig. 10.

Composite temperature anomalies for (a) high- and (b) low-VIWVIPWP years; (c),(d) as in (a) and (b), respectively, but for precipitation. The anomaly values represent the difference between the high- or low-VIWVIPWP years and the normal-VIWVIPWP years.

By calculating the spatial correlations, we also detected the linkages among VIWVIPWP, VIWVglobal, and precipitation on a global scale (Fig. 11). Figure 11a shows that most land in the NH has a positive significant correlation at the 95% confidence level, but negative significant correlation at the 95% confidence level in the SH. Figure 11b shows that a strong positive correlation between VIWVIPWP and precipitation is associated with the following regions: most of Eurasia, eastern North America, and central Africa; a strong negative significant correlation occurs for central southern Africa and South America. These results demonstrate a close connection between VIWVIPWP and VIWVglobal and show that changes in precipitation are closely related to changes in VIWVIPWP. We also investigated the zonal changes in VIWVglobal, precipitation, and NDVI associated with VIWVIPWP anomalies (Fig. 12). The VIWVglobal, precipitation, and NDVI values during high-VIWVIPWP years were obviously higher than those during low-VIWVIPWP years, and these differences were considerable. For example, during high-VIWVIPWP years, the NDVI was obviously higher than normal across all latitudes (Fig. 12c), with VIWVglobal, temperature, and precipitation showing similar trends, and vice versa. Figure 12 clearly demonstrates these results. In addition, we also analyzed the relationships between the NDVI and the temperature, precipitation, and VIWVglobal during high (low) VIWVIPWP years (Table 3). Notably, in both high- and low-VIWVIPWP years, VIWVglobal had the strongest relationship with the NDVI among all the studied variables. This result implies that NDVI anomalies are closely linked with VIWVglobal.

Fig. 11.

Spatial correlation coefficient between VIWVIPWP and the (a) VIWV and (b) NDVI during AMJ from 1982 to 2015. Coefficients greater than +0.2 or less than −0.2 are significant at the 95% confidence level.

Fig. 11.

Spatial correlation coefficient between VIWVIPWP and the (a) VIWV and (b) NDVI during AMJ from 1982 to 2015. Coefficients greater than +0.2 or less than −0.2 are significant at the 95% confidence level.

Fig. 12.

Composite zonal distribution of (a) VIWVglobal, (b) precipitation, and (c) NDVI anomalies for high-VIWVIPWP years (black lines) and low-VIWVIPWP years (red lines). The anomaly values represent the difference between the high- or low-VIWVIPWP years and normal-VIWVIPWP years.

Fig. 12.

Composite zonal distribution of (a) VIWVglobal, (b) precipitation, and (c) NDVI anomalies for high-VIWVIPWP years (black lines) and low-VIWVIPWP years (red lines). The anomaly values represent the difference between the high- or low-VIWVIPWP years and normal-VIWVIPWP years.

Table 3.

The relationship between zonal NDVI, temperature, precipitation, and VIWVglobal during VIWVIPWP anomaly years. Plus and minus signs indicate high- and low-VIWVIPWP years, respectively; two asterisks indicate a strong relationship (correlation at the 99% level, estimated by a local Student’s t test).

The relationship between zonal NDVI, temperature, precipitation, and VIWVglobal during VIWVIPWP anomaly years. Plus and minus signs indicate high- and low-VIWVIPWP years, respectively; two asterisks indicate a strong relationship (correlation at the 99% level, estimated by a local Student’s t test).
The relationship between zonal NDVI, temperature, precipitation, and VIWVglobal during VIWVIPWP anomaly years. Plus and minus signs indicate high- and low-VIWVIPWP years, respectively; two asterisks indicate a strong relationship (correlation at the 99% level, estimated by a local Student’s t test).

e. The large-scale circulation features associated with VIWVIPWP anomalies

To reveal the mechanism underlying the response of the NDVI to VIWVIPWP anomalies, we conducted a composite analysis to investigate the atmospheric circulation climatological patterns associated with VIWVIPWP anomalies (Fig. 13). No matter how high or low the VIWVIPWP was, the SLP anomaly pattern during high-VIWVIPWP years (Figs. 13a,b) revealed that strong cyclonic (low pressure) activity covered more land in the NH, whereas more land in the SH was controlled by strong anticyclonic (high pressure) activity. In the meridional direction, there was strong anticyclonic activity in the Pacific Ocean and Australia but cyclonic activity covering the IPWP region. Compared with the SLP anomaly pattern during low-VIWVIPWP years (Fig. 13b), the pattern was enhanced in high-VIWVIPWP years, and areas with low-induced SLP were expanded (Fig. 13a).

Fig. 13.

Composite SLP (shading and contour lines; hPa) for (a) high- and (b) low-VIWVIPWP years; (c),(d) as in (a) and (b), respectively, but for VIWVglobal instead of VIWVIPWP (shading; kg m−2) and uυ wind at 850 hPa (vectors; m s−1).

Fig. 13.

Composite SLP (shading and contour lines; hPa) for (a) high- and (b) low-VIWVIPWP years; (c),(d) as in (a) and (b), respectively, but for VIWVglobal instead of VIWVIPWP (shading; kg m−2) and uυ wind at 850 hPa (vectors; m s−1).

During high-VIWVIPWP years, the southwestward Eurasia trough (low pressure), the Pacific quasi-south–north ridge (high pressure), the North American northeastward ridge, and the South American quasi-south–north weak ridge form a specific global atmospheric circulation mode; in addition, the VIWV over most of the land and the mid- and high-latitude oceans is extremely high, and the high value center of global VIWV is located in the Pacific Ocean in the equatorial region. The Pacific quasi-south–north ridge causes the anomalous southeast wind to bring VIWV originating from the Pacific Ocean to most of eastern Eurasia; the high-latitude northwestward ridge causes the anomalous westerly winds to transport VIWV from the Arctic Ocean into northwestern Eurasia. Underlying South Africa’s weak high pressure and the Pacific quasi-south–north ridge, the VIWV from the Mediterranean and the Atlantic Ocean is heading north; the South American ridge causes the anomalously enhanced southwesterly wind to transport the Atlantic VIWV to southern Africa. Because the VIWV from the Pacific Ocean is blocked by the Cordillera Mountains, making it difficult to reach the central and eastern parts of the Americas, only the anomalous north wind brings the Arctic Ocean water vapor to the areas beyond the eastern border of the Cordillera Mountains in North America. The anomalous northeast wind will bring VIWV from the Atlantic Ocean into the northern part of South America. Because the VIWV from the Atlantic Ocean has a suppressed weak high pressure and abnormal westerly winds, the VIWV over the south central region of southern Africa declined.

Compared with the atmospheric circulation and VIWV conditions during high-VIWVIPWP years, during low-VIWVIPWP years the high-latitude ridges formed in the NH, the Eurasian southwestward trough, the Pacific quasi-south–north ridge, the North American northeastward ridge, and the South American quasi-south–north ridge form a relatively weak, specific global atmospheric circulation mode. At the same time, the VIWV over most land areas (except the African region) and the ocean in the NH is unusually reduced. The high-value center of VIWV is also shifted southward from the equatorial region. In the context of a VIWV decline in the NH, the anomalous north wind led by the strong high-latitude ridge suppressed the movement of Pacific water vapor to Eurasia; the anomalous west wind prevented VIWV from the Atlantic from being transported to North America. Because Australia’s high pressure has hindered the westward transmission of water vapor from the eastern Pacific, Australia’s water vapor value is lower. In addition, the African water vapor high value center has a shift eastward in the northwestern part of North Africa (during high-VIWVIPWP years), indicating that the Atlantic water vapor has compensated for the shortage of water vapor in Africa with an anomalously enhanced westerly transport.

The anomalous patterns of VIWVglobal show that during high-VIWVIPWP years, a remarkable increase in VIWV over land appeared, with the opposite trend in low-VIWVIPWP years, excluding central Africa and the southern parts of North America. These patterns were similar to the precipitation anomaly patterns (Figs. 10c,d). The anomaly patterns of the u and υ wind components at 850 hPa revealed that during high-VIWVIPWP years (Figs. 13c,d) a strong interaction between enhanced cyclonic activity over the land and anticyclonic activity over the Pacific Ocean brought more VIWV from the ocean to the land by the enhanced wind. When VIWVIPWP was low, a weak interaction between the subdued cyclonic activity over the land and subdued anticyclonic activity over the Pacific Ocean brought less atmospheric water vapor from the ocean to the land by the subdued wind.

To further verify these results, we depicted the zonal variation in VIWVglobal and u–υ wind at 850 hPa over the two years (Fig. 14). Figure 14 shows that changes in VIWV over land are closely related to anomalous wind field changes. During high-VIWVIPWP years, the VIWV in the NH is obviously higher than during normal- and low-VIWV years. The low-latitude region (0°–30°N) has an anomalous southeast wind, an enhanced anomalous east wind in the midlatitude region, and an enhanced anomalous northwest wind in the high-latitude region. In the SH, the VIWV in the midlatitude region is much higher than that during normal- and low-VIWVIPWP years, related to anomalously enhanced southwesterly winds. The VIWV of low-latitude regions is lower than that during the low years but higher than that of the normal years, related to the anomalously strong south winds. The high-latitude VIWV is higher than that of normal and low years, and the north winds are enhanced anomalously. During low-VIWVIPWP years, the VIWV in the NH is significantly lower than that during normal-VIWVIPWP years, the northerly winds in the low latitudes are anomalously enhanced, the northwest winds in the midlatitude areas are anomalous, and the VIWV in the low-latitude regions of the SH is anomalously increased, which is related to the anomalously enhanced north wind and the Pacific water vapor high-value centers move southward. In addition, the opposite wind direction for most latitudes is related to the difference in global atmospheric circulation modes during low- and high-VIWVIPWP years.

Fig. 14.

Composite zonal distribution of VIWV (lines) and uυ wind (vectors; m s−1) at 850 hPa. Plus and minus signs indicate high- and low-VIWVIPWP years, respectively.

Fig. 14.

Composite zonal distribution of VIWV (lines) and uυ wind (vectors; m s−1) at 850 hPa. Plus and minus signs indicate high- and low-VIWVIPWP years, respectively.

The results show that during high-VIWVIPWP years, a strong interaction led to an increased atmospheric water vapor flux brought by the anomalously enhanced southeast wind in 30°S–30°N. To a lesser extent, the anomalous northeast wind inhibited the transport of atmospheric water vapor from the ocean to the land.

4. Conclusions and discussion

Previously, many studies have indicated that global vegetation is affected by SST, but relatively little attention has been paid to the influence of VIWV over a key region of the ocean on terrestrial vegetation growth activity, despite general awareness that VIWV originating from the ocean affects the spatial distribution of precipitation through LACs. Based on gridded SST, CRU, ERA-Interim, and NDVI data from 1982 to 2015, the impacts of VIWVIPWP anomalies on global vegetation growth during AMJ are investigated in detail. A strong correlation between VIWVIPWP and the global NDVI during AMJ suggests that VIWVIPWP is an important factor affecting TVG. The dynamic processes linking global NDVI to VIWVIPWP anomalies are then explored. We found a positive (or negative) response of global vegetation to an increase (or decrease) in the VIWVIPWP anomaly. For example, during high-VIWVIPWP years, the NH and SH NDVI exhibit an obvious increase of 9% and 2%, respectively, but a 4% and 3% decrease during low-VIWVIPWP years. Correspondingly, during high-VIWVIPWP years, an increase in temperature and precipitation occurs over most of the global land area, while the opposite occurs during low-VIWVIPWP years. In addition, the strong spatial correlation between VIWVIPWP and NDVI implies that VIWVIPWP affects the spatial distribution of precipitation. Moreover, a stronger positive correlation between zonal-averaged NDVI and VIWV than that with temperature and precipitation also suggests that the VIWV is an important factor affecting the variation in NDVI. The results of the composite analysis of LAC features associated with VIWVIPWP show that during high-VIWVIPWP years, there is an enhanced interaction between one Pacific ridge and two land troughs, resulting in a strong anomalous southeast (or southwest) wind; therefore, more VIWV originating from the ocean is brought to land, and the increase in terrestrial atmosphere water vapor results in more favorable conditions for precipitation, enhancing the terrestrial precipitation probability. Because of the anomaly-enhanced interactions of IPWP atmospheric water vapor with the atmosphere and ocean during high-VIWVIPWP years, more and warmer water vapor is transported to the land, thus improving global vegetation growth. In addition, the data showed that a variety of VIWVIPWP anomalies occur, and LAC-induced atmospheric water vapor originating from the ocean is transported to the land, which creates favorable conditions for the formation of precipitation over land; thus, the AMJ periods with more precipitation and warmth improve vegetation growth over land, globally. Notably, a new understanding of the effects of ocean activities on the terrestrial ecosystem is depicted in this scientific, fact-based study, which shows that global vegetation growth activity anomalies can be induced by the VIWVIPWP. Surprisingly, VIWVIPWP can be a new and useful index tool for monitoring and predicting global vegetation growth change.

However, this paper mainly focuses on the impacts of VIWVIPWP anomalies on global vegetation growth during AMJ. Hence, during other seasons and at an interannual scale, the possibly different response mechanisms of global vegetation growth to VIWVIPWP anomalies should be further investigated. In addition, because of the short times series of the remote sensing data, a further investigation and verification of the response mechanisms of global vegetation growth to VIWVIPWP anomalies should be conducted by using a longer time series of remote sensing data. Furthermore, a more detailed understanding of these response mechanisms by simulating these processes with a state-of-art ocean–land model is needed.

Finally, in the near future with global warming and more anthropogenic activities, a continuing increase in SST and terrestrial temperature may increase evaporation and VIWV but decrease convective clouds (De Deckker 2016). If SSTs are ≥31.05°C, the high convective clouds may vanish (Waliser and Graham 1993; Waliser et al. 1993; Webster et al. 1998; Webster 1994). Under such higher temperature conditions, cloud cover will be significantly reduced, causing longwave radiation to increase, forcing a decline in the nocturnal temperature. In addition, the thermal difference between sea and land is weakened, resulting in a decrease in the land–sea pressure difference, which weakens the LAC interaction between land and ocean, eventually causing less moisture transport from the ocean to the land. In particular, the reduced cloud cover does not favor the formation of precipitation, eventually causing a decline in precipitation. In summary, as a result of such climate conditions, drought would significantly increase in the future, eventually causing vegetation degradation. The reduction in nocturnal temperature also would inhibit terrestrial vegetation growth activity.

Acknowledgments

We thank three anonymous reviewers and Dr. Jason Evans (the editor) for their comments and suggestions, which helped improve the manuscript. This research was supported by the National Key R&D Program of China (2017YFC0503905) and the National Natural Science Foundation of China (Grant 41801082). The authors declare that they have no competing interests. We thank Mrs. Na Zhao (Xi’an, Shaanxi, China) for assisting in color matching for images.

REFERENCES

REFERENCES
Annamalai
,
H.
,
J.
Hafner
,
K.
Sooraj
, and
P.
Pillai
,
2013
:
Global warming shifts the monsoon circulation, drying South Asia
.
J. Climate
,
26
,
2701
2718
, https://doi.org/10.1175/JCLI-D-12-00208.1.
D’Arrigo
,
R.
, and Coauthors
,
2006
:
The reconstructed Indonesian warm pool sea surface temperatures from tree rings and corals: Linkages to Asian monsoon drought and El Niño–Southern Oscillation
.
Paleoceanography
,
21
,
PA3005
, https://doi.org/10.1029/2005PA001256.
De Deckker
,
P.
,
2016
:
The Indo-Pacific warm pool: Critical to world oceanography and world climate
.
Geosci. Lett.
,
3
,
20
, https://doi.org/10.1186/s40562-016-0054-3.
Dee
,
D. P.
, and Coauthors
,
2011
:
The ERA-Interim reanalysis: Configuration and performance of the data assimilation system
.
Quart. J. Roy. Meteor. Soc.
,
137
,
553
597
, https://doi.org/10.1002/qj.828.
Fang
,
J.
,
S.
Piao
,
J.
He
, and
W.
Ma
,
2004
:
Increasing terrestrial vegetation activity in China, 1982–1999
.
Sci. China
,
47C
,
229
240
.
Feng
,
J.
, and
J.
Li
,
2013
:
Contrasting impacts of two types of ENSO on the boreal spring Hadley circulation
.
J. Climate
,
26
,
4773
4789
, https://doi.org/10.1175/JCLI-D-12-00298.1.
Feng
,
R.
,
J.
Li
, and
J.
Wang
,
2011
:
Regime change of the boreal summer Hadley circulation and its connection with the tropical SST
.
J. Climate
,
24
,
3867
3877
, https://doi.org/10.1175/2011JCLI3959.1.
Funk
,
C.
,
A.
Hoell
,
S.
Shukla
,
I.
Bladé
,
B.
Liebmann
,
J. B.
Roberts
,
F. R.
Robertson
, and
G.
Husak
,
2014
:
Predicting East African spring droughts using Pacific and Indian Ocean sea surface temperature indices
.
Hydrol. Earth Syst. Sci.
,
18
,
4965
4978
, https://doi.org/10.5194/hess-18-4965-2014.
Hirahara
,
S.
,
M.
Ishii
, and
Y.
Fukuda
,
2014
:
Centennial-scale sea surface temperature analysis and its uncertainty
.
J. Climate
,
27
,
57
75
, https://doi.org/10.1175/JCLI-D-12-00837.1.
Holben
,
B. N.
,
1986
:
Characteristics of maximum-value composite images from temporal AVHRR data
.
Int. J. Remote Sens.
,
7
,
1417
1434
, https://doi.org/10.1080/01431168608948945.
Huang
,
B.
, and
V. M.
Mehta
,
2004
:
Response of the Indo-Pacific warm pool to interannual variations in net atmospheric freshwater
.
J. Geophys. Res. Oceans
,
109
,
C06022
, https://doi.org/10.1029/2003JC002114.
Huang
,
R.
, and
F.
Sun
,
1992
:
Impacts of the tropical western Pacific on the East Asian summer monsoon
.
J. Meteor. Soc. Japan
,
70
,
243
256
.
Huete
,
A. R.
,
1988
:
A soil-adjusted vegetation index (SAVI)
.
Remote Sens. Environ.
,
25
,
295
309
, https://doi.org/10.1016/0034-4257(88)90106-X.
Jones
,
P.
, and
I.
Harris
,
2013
: CRU TS3. 21: Climatic Research Unit (CRU) time-series (TS) version 3.21 of high resolution gridded data of month-by-month variation in climate (Jan. 1901–Dec. 2012). NCAS British Atmospheric Data Centre, https://doi.org/10.5285/D0E1585D-3417-485F-87AE-4FCECF10A992.
Li
,
Y.
,
J.
Li
, and
J.
Feng
,
2013
:
Boreal summer convection oscillation over the Indo-western Pacific and its relationship with the East Asian summer monsoon
.
Atmos. Sci. Lett.
,
14
,
66
71
, https://doi.org/10.1002/asl2.418.
Ma
,
J.
, and
J.
Li
,
2008
:
The principal modes of variability of the boreal winter Hadley cell
.
Geophys. Res. Lett.
,
35
,
L01808
, https://doi.org/10.1029/2007GL031883.
Neale
,
R.
, and
J.
Slingo
,
2003
:
The Maritime Continent and its role in the global climate: A GCM study
.
J. Climate
,
16
,
834
848
, https://doi.org/10.1175/1520-0442(2003)016<0834:TMCAIR>2.0.CO;2.
Niedermeyer
,
E. M.
,
A. L.
Sessions
,
S. J.
Feakins
, and
M.
Mohtadi
,
2014
:
Hydroclimate of the western Indo-Pacific warm pool during the past 24,000 years
.
Proc. Natl. Acad. Sci. USA
,
111
,
9402
9406
, https://doi.org/10.1073/pnas.1323585111.
Russell
,
J. M.
, and Coauthors
,
2014
:
Glacial forcing of central Indonesian hydroclimate since 60,000 y B.P
.
Proc. Natl. Acad. Sci. USA
,
111
,
5100
5105
, https://doi.org/10.1073/pnas.1402373111.
Sardeshmukh
,
P. D.
, and
B. J.
Hoskins
,
1988
:
The generation of global rotational flow by steady idealized tropical divergence
.
J. Atmos. Sci.
,
45
,
1228
1251
, https://doi.org/10.1175/1520-0469(1988)045<1228:TGOGRF>2.0.CO;2.
Shan
,
H.
,
Y.
Guan
, and
J.
Huang
,
2014
:
Surface air temperature patterns on a decadal scale in china using self-organizing map and their relationship to Indo-Pacific warm pool
.
Int. J. Climatol.
,
34
,
3752
3765
, https://doi.org/10.1002/joc.3943.
Shin
,
S.-I.
,
P. D.
Sardeshmukh
, and
S.-W.
Yeh
,
2011
:
Sensitivity of the northeast Asian summer monsoon to tropical sea surface temperatures
.
Geophys. Res. Lett.
,
38
,
L22702
, https://doi.org/10.1029/2011GL049391.
Simmons
,
A.
,
2006
: ERA-Interim: New ECMWF reanalysis products from 1989 onwards. ECMWF Newsletter, No. 110, ECMWF, Reading, United Kingdom, 25–36.
U.S. Defense Mapping Agency
,
1987
: Supplement to Department of Defense World Geodetic System 1984 Technical Report, Vol. 2. DMA, 169 pp.
Visser
,
K.
,
R.
Thunell
, and
L.
Stott
,
2003
:
Magnitude and timing of temperature change in the Indo-Pacific warm pool during deglaciation
.
Nature
,
421
,
152
, https://doi.org/10.1038/nature01297.
Waliser
,
D. E.
, and
N. E.
Graham
,
1993
:
Convective cloud systems and warm-pool sea surface temperatures: Coupled interactions and self-regulation
.
J. Geophys. Res.
,
98
,
12 881
12 893
, https://doi.org/10.1029/93JD00872.
Waliser
,
D. E.
,
N. E.
Graham
, and
C.
Gautier
,
1993
:
Comparison of the highly reflective cloud and outgoing longwave radiation datasets for use in estimating tropical deep convection
.
J. Climate
,
6
,
331
353
, https://doi.org/10.1175/1520-0442(1993)006<0331:COTHRC>2.0.CO;2.
Webster
,
P. J.
,
1994
:
The role of hydrological processes in ocean–atmosphere interactions
.
Rev. Geophys.
,
32
,
427
476
, https://doi.org/10.1029/94RG01873.
Webster
,
P. J.
,
V. O.
Magana
,
T.
Palmer
,
J.
Shukla
,
R.
Tomas
,
M.
Yanai
, and
T.
Yasunari
,
1998
:
Monsoons: Processes, predictability, and the prospects for prediction
.
J. Geophys. Res.
,
103
,
14 451
14 510
, https://doi.org/10.1029/97JC02719.
Zhou
,
L.
,
C. J.
Tucker
,
R. K.
Kaufmann
,
D.
Slayback
,
N. V.
Shabanov
, and
R. B.
Myneni
,
2001
:
Variations in northern vegetation activity inferred from satellite data of vegetation index during 1981 to 1999
.
J. Geophys. Res.
,
106
,
20 069
20 083
, https://doi.org/10.1029/2000JD000115.
Zhou
,
L.
,
R.
Kaufmann
,
Y.
Tian
,
R.
Myneni
, and
C.
Tucker
,
2003
:
Relation between interannual variations in satellite measures of northern forest greenness and climate between 1982 and 1999
.
J. Geophys. Res.
,
108
,
4004
, https://doi.org/10.1029/2002JD002510.
Zhou
,
X.
,
J.
Li
,
F.
Xie
,
R.
Ding
,
Y.
Li
,
S.
Zhao
,
J.
Zhang
, and
Y.
Li
,
2018
:
The effects of the Indo-Pacific warm pool on the stratosphere
.
Climate Dyn.
,
51
,
4043
4064
, https://doi.org/10.1007/s00382-017-3584-2x.
Zhu
,
M.
,
L.
Stott
,
B.
Buckley
,
K.
Yoshimura
, and
K.
Ra
,
2012
:
Indo–Pacific warm pool convection and ENSO since 1867 derived from Cambodian pine tree cellulose oxygen isotopes
.
J. Geophys. Res.
,
117
,
D11307
, https://doi.org/10.1029/2011JD017198.

Footnotes

g

Current affiliation: Key Laboratory of Ecosystem Network Observation and Modeling, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing, China.

© 2019 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).