Increased Browning of Woody Vegetation due to Continuous Seasonal Droughts in Yunnan Province, China

CHEN Hong-Ping , JIA Gen-Suo FENG Jin-Ming and DONG Yan-Sheng

1 Key Laboratory of Regional Climate-Environment for East Asia, Institute of Atmospheric Physics, Chinese Academy of Sciences,Beijing 100029, China

2 University of Chinese Academy of Sciences, Beijing 100049, China

3 Beijing Research Center for Information Technology in Agriculture, Beijing 100097, China

1 Introduction

Drought is a climate event, characterized by a longer than normal period of abnormally low precipitation and water deficit (Dai, 2011). To human societies and ecosystems, it is a serious and complex climate event (Wilhite,1993). In the context of global warming, persistent drought may increase (The Intergovernmental Panel on Climate Change (IPCC), 2007), and may put great pressure on various ecosystems. The response of forest ecosystems to drought has become a hotspot of climate change research (Saleska et al., 2007; Malhi et al., 2008;Phillips et al., 2009; Brando et al., 2010; Xu et al., 2011;Davidson et al., 2012). In addition, when monitoring global vegetation photosynthetic effectiveness (Huete et al., 2002; Justice et al., 2002) and photosynthetic carbon fixation (Huete et al., 2006; Yang et al., 2007; Brando et al., 2010), the enhanced vegetation index (EVI) is widely used.

Located in southwest China, Yunnan Province has a tropical and subtropical climate, distinct wet and dry seasons, dense forests, and rich biodiversity. In this region,85% of rainfall occurs in the wet season from April to September. From October 2009 to March 2010, Yunnan Province suffered from the severest drought since 1951.Previous studies have mainly focused on analyzing the cause of the 2010 drought (e.g., Barriopedro et al., 2012)and its impact on water resources, agricultural production,and primary productivity (e.g., Piao et al., 2010; Barriopedro et al., 2012; Zhang et al., 2012). Drought continued in the 2011 wet season and the 2012 dry season (from October 2011 to March 2012), but little attention has been paid to the impact of these continuous seasonal droughts on the vegetation over Yunnan Province.

In the paper, we use the Moderate Resolution Imaging Spectroradiometer (MODIS) EVI greenness index to analyze the impacts of these seasonal droughts on the woody ecosystem. The objectives were to find out whether the droughts would exacerbate the decline of woody vegetation greenness, and which woody vegetation type was more sensitive to drought than other types.

2 Data and method

2.1 Data

2.1.1 Meteorological stations data

The monthly rainfall data from 33 meteorological stations over Yunnan Province from April 1951 to May 2012 were provided by the China Meteorological Data Sharing Service network (http://cdc.cma.gov.cn/home.do).

2.1.2 Satellite precipitation

The version 7 Tropical Rainfall Measuring Mission(TRMM) monthly precipitation products (3B43) (http://trmm.gsfc.nasa.gov) were used. The data are at a spatial resolution of 0.25° × 0.25° and are from January 1998 to June 2012.

2.1.3 Vegetation index

The Collection 5 (C5) MOD13A2 EVI data provided by MODIS (https://lpdaac.usgs.gov) were used. The data are at a spatial resolution of 1 km × 1 km, a temporal resolution of 16 days, and are from January 2001 to June 2012.

以及邊界條件 其中, L≡(Dy+Dz-, Dy=?y, Dz=?z. 線性穩定性方程的特征值σ=σr+iσi可以用于研究流動系統的穩定性: σr是角頻率, 而σi是線性增長率. 對于任意實波數k, 如果存在某個特征模態的σi>0, 那么流動將是線性時間不穩定的.

2.1.4 Land cover

The 2009 European Space Agency (ESA) global land cover data (Globcover) (http://ionia1.esrin.esa.int/), which is at a spatial resolution of 300 m × 300 m, were used.After re-sampling the data into a 1 km × 1 km spatial resolution using the nearest method, we converted the woody vegetation based on the International Geosphere Biosphere Program (IGBP) land cover classification criteria as Herold et al. (2008) did. The whole woody vegetation accounts for 57.95% of the entire Yunnan Province area, including evergreen needleleaf forest (ENF), evergreen broadleaf forest (EBF), deciduous broadleaf forest(DBF), mixed forest (MF), and shrubland (SL) (Fig. 1).Of this woody vegetation, the various types occupy the following amounts of the whole: ENF 27.00%, EBF 23.24%, SL 35.63%, DBF 8.42%, and MF 5.72%.

2.2 Methods

We calculated the standardized anomalies of precipitation and EVI pixel-by-pixel for 2010, 2011, and 2012 using

Figure 1 Aggregated International Geosphere Bioshere Program(IGBP) woody vegetation classes of Yunnan Province at 1 km × 1 km spatial resolution.

where φ is the standardized anomaly of a given quality(precipitation or EVI) calculated from its value, x, in the same year and the long-term mean, μ, standard deviation,and σ, over a reference period. If φ is less than ?1, then the area is drought-stricken (Saleska et al., 2007; Xu et al., 2011) or there is vegetation browning (Samanta et al.,2010).

2.2.1 Precipitation standardized anomaly

Because there were dry season droughts in 2010 and 2012, the dry season precipitation standardized anomaly was calculated. The dry season is from October to March.The precipitation standardized anomaly is calculated using only valid values. The monthly precipitation value of a pixel is considered valid if the value is not equal to?9999. x is the accumulation of valid TRMM monthly precipitation values in the dry season, and μ and σ are the mean and standard deviation of valid precipitation values in the dry seasons from 1998 to 2009.

Because through October to December, there is snow over Northwest Yunnan Province, the January to March(JFM) EVI standardized anomaly was calculated. The EVI standardized anomaly is calculated using only valid values. The validness of a pixel EVI value is determined by the combined rules set by Xu et al. (2011). x is the mean of valid EVI values in JFM, and μ and σ are the mean and standard deviation of valid EVI values in JFM from 2001 to 2009.

The 2010 and 2012 droughts went through January to March, considering the lagged responses of vegetation to drought-stress (Kileshye Onema and Taigbenu, 2009; Allen et al., 2010), thus besides JFM, the post-drought months from April to June (AMJ) were also considered to describe the responses of different woody vegetation types to droughts.

3 Results

3.1 Droughts impact in dry season

Figure 2 Station-averaged dry season and wet season rainfall anomaly percentages from 1951 to 2012 over Yunnan Province. Dry season is from October to March. Wet season is from April to September. Wet season rainfall anomaly percentages were ?0.06%, ?0.43%, and ?0.22% in 1995,2006, and 2010, so they could not be shown. Dry season rainfall anomaly percentage in 1951 was zero, because monthly rainfall data was from April 1951.

From the monthly rainfall data of 33 meteorological stations over Yunnan Province, the rainfall deficit in the 2010 dry season was ?60.98%. This was a major deficit compared to the ?58.31%, ?53.39%, and ?57.48% in the dry seasons of 1963, 1969, and 1979. Drought in the 2010 dry season was thus the most severe one (Fig. 2). Mean-while, there were ?15.05% and ?31.88% rainfall deficits in the 2011 wet season and the 2012 dry season, giving three successive years of droughts. The rainfall deficit in the 2010 dry season was almost twice that in 2012. Based on the TRMM precipitation standardized anomaly, almost the entire province was affected by drought in 2010,compared with nearly no drought-stricken areas in 2011 and mainly northwest, northern, and eastern parts in 2012(Figs. 3a1–c1). In 2010, 85.66% of the total area suffered from drought (372710 km2) compared to 48.50% in 2012(211000 km2). Thus the 2010 drought impacted an area 1.77 times larger than the 2012 drought.

On the basis of the EVI standardized anomaly, the woody vegetation of western and Southeast Yunnan Province showed browning in 2010, while in 2012, the browning of the vegetation was in Northwest and eastern Yunnan Province (Figs. 3a2 and 3c2). Most parts of these spatial distributions were consistent with the spatial distributions of the precipitation standardized anomalies in the 2010 and 2012 dry seasons. There were also areas where severe drought occurred, but there was only slight browning, for example, Northwest Yunnan Province in 2010 (Figs. 3a1 and 3a2). In 2011, a small portion of woody vegetation in the Southeast and Northwest Yunnan Province showed browning (Fig. 3b2). In 2010 and 2012,the EVI anomalies of all the woody vegetation displayed positively skewed frequency distributions (Fig. 4a1). In 2010, 31.47% of the woody vegetation (65028 km2)showed browning, compared to 28.08% in 2012 (57959 km2) (Table 1). Thus, although the 2010 drought was much more severe than the 2012 drought, the browning area of woody vegetation in JFM 2012 was slightly smaller than that in JFM 2010.

We divided the woody vegetation in Yunnan Province into two groups to present their different sensitivities to drought. ENF, EBF, and SL were in the first group; DBF and MF were in the second group. The EVI anomalies of all the different woody vegetation types displayed positively skewed frequency distributions in JFM 2010 and 2012 (Figs. 4a2 and 4c2). When compared with the browning percentages in JFM 2010, only ENF and EBF show increased browning percentages in JFM 2012 (Table 1), indicating that the extreme drought in 2010 followed by a moderate drought in 2012 may have caused an accumulative impact on these two forest types over Yunnan Province.

3.2 Impact of droughts in the post-drought months

Northwest, Southwest, and Southeast Yunnan Province suffered from drought in AMJ 2010, unlike northern and Southeast Yunnan Province in 2011 and Northwest and southern Yunnan Province in 2012 (Figs. 3d1–f1). In 2010, 54.51% of the Yunnan Province area (237180 km2),compared with 53.63% and 41.42% in 2011 (233350 km2)and 2012 (180220 km2), was affected by drought. The drought in AMJ 2010 impacted an area 1.32 times larger than that in AMJ 2012.

Woody vegetation showed browning mostly in Northwest, northern, and Southeast Yunnan Province in AMJ 2010 and 2012 (Figs. 3d2 and 3f2). Most parts of these spatial distributions were consistent with the spatial distributions of the precipitation standardized anomalies in the same time. There were also areas where positive precipitation anomalies occurred, but browning was evident,for example, Southeast Yunnan Province in AMJ 2012(Figs. 3f1 and 3f2). Almost no woody vegetation became brown in AMJ 2011 (Fig. 3e2), despite 53.63% of the area in Yunnan Province (233350 km2) being under drought in the same time (Fig. 3e2). The EVI anomaly frequency distributions of the whole woody vegetation and different woody vegetation types in AMJ 2010 and 2012 were all positively skewed, but were all negatively skewed in AMJ 2011 (Figs. 4b1 and 4d2–f2). During AMJ 2010, 35.77%of the whole woody vegetation (73023 km2) showed browning compared with 39.63% in AMJ 2012 (80822 km2) (Table 1). Thus a moderate 2012 drought following the extreme 2010 drought amplified the browning of woody vegetation over Yunnan Province.

Table 1 EVI standard anomalies (φ) of different woody vegetation types and the whole woody vegetation in January-February-March (JMF) and April-May-June (AMJ) 2010, 2011, and 2012 over Yunnan Province. φ≤?1 means woody vegetation browning; “ENF” is evergreen needleleaf forest; “EBF” is evergreen broadleaf forest; “SL” is shrubland; “DBF” is deciduous broadleaf forest; “MF” is mixed forest, and “ALL” is the whole woody vegetation.

Figure 3 Spatial patterns of TRMM precipitation standardized anomalies in (a1) dry season 2010; (b1) dry season 2011; (c1) dry season 2012; (d1)AMJ 2010; (e1) AMJ 2011, and (f1) AMJ 2012. Dry season was from October of the previous year to March of the next year. Spatial patterns of Enhanced Vegetation Index (EVI) standardized anomalies in (a2) JFM 2010; (b2) JFM 2011; (c2) JFM 2012; (d2) AMJ 2010; (e2) AMJ 2011; (f2) AMJ 2012.

More ENF (40.27% and 43.38%, 23193 km2and 24967 km2) and more MF (49.56% and 55.36%, 5294 km2and 5910 km2) showed browning during AMJ 2010 and 2012 in the first and second woody vegetation group(Table 1). The correlations between the TRMM precipitation standardized anomaly and the EVI standardized anomalies of ENF and EBF were analyzed taking into account the different spatial resolutions of the TRMM precipitation and the MODIS EVI. Firstly, setting that if in one TRMM precipitation pixel, there was more than 50 ENF or EBF pixels, then, the TRMM precipitation pixel is a valid TRMM precipitation pixel. Secondly, calculating the mean EVI standardized anomaly of all the ENF or EBF pixels that within the valid TRMM precipitation pixel and then analyzing the correlations between it and the standardized anomaly of the valid TRMM precipitation pixel. The correlation coefficients (r) between the TRMM precipitation standardized anomaly and the EVI standardized anomalies of ENF and EBF were 0.16 and 0.11, respectively, in AMJ 2012. Both the correlations were through the significant test (p < 0.1). Thus, from the browning percentage and the relativity with the TRMM precipitation standardized anomaly, ENF was more sensitive to drought than other woody vegetation types.

4 Discussion and conclusion

4.1 The importance of precipitation

Figure 4 Frequency distributions of EVI standardized anomalies of the whole woody vegetation in (a1) JFM 2010, 2011, and 2012 and (b1) AMJ 2010, 2011, and 2012. Frequency distributions of EVI standardized anomalies of different woody vegetation types in (a2) JFM 2010; (b2) JFM 2011;(c2) JFM 2012; (d2) AMJ 2010; (e2) AMJ 2011; (f2) AMJ 2012.

In 2010 and 2012, most of the spatial distributions between woody vegetation browning and the TRMM precipitation anomalies were consistent. During JFM 2011,woody vegetation browned, despite there being almost no drought. In addition to the uneven spatial distribution of precipitation, other possible reasons may include (1) insufficient precipitation to replenish the serious depletion of plant-available soil water (PAW) (Nepstad et al., 2007)caused by the 2010 severest dry season drought, and (2)the lagged responses of woody vegetation to drought.During AMJ 2011, the spatial distribution of precipitation was also uneven and the response of woody vegetation to drought also lagged, besides, although the amount of precipitation reduced through April to June 2011, the absolute amount of precipitation could supply the woody vegetation. So, in AMJ 2011, although almost droughtaffected, large parts of the Yunnan Province showed greening.

4.2 Three-year drought accumulative effect on woody vegetation

The three-year continuous seasonal droughts intensified woody vegetation browning. Although the drought was not as severe in the 2012 dry season as it was in 2010, there were larger browning areas in 2012 than in 2010 through January to June over Yunnan Province. In addition to the accumulative effect of droughts on woody vegetation, other possible reasons which could make wood vegetation show browning include tree mortality(Betts et al., 2008), fires (Lewis et al., 2011), and inspects(Brando et al., 2010). These reasons might combine with each other, and have an accumulative effect on woody vegetation browning, so they need to be further validated through field investigation in future work.

4.3 The response of woody vegetation to droughts

Considering the lagged response of woody vegetation to drought, we discuss the response in the post-drought months (April to June). We find that in the first and second woody vegetation group, ENF and MF showed a larger browning percentage in both AMJ 2010 and 2012.The possible reasons why ENF has a larger browning percentage are (1) ENF is almost in the drought-stricken regions in AMJ 2010 and 2012; (2) ENF is distributed mainly over Northwest Yunnan Province, and because of the monsoon influence, there is comparatively less rainfall in Northwest than in South Yunnan Province; (3) Northwest Yunnan Province, where ENF is mainly distributed,is upstream of rivers. In the second group, MF is the mixture of ENF and DBF, the larger browning percentage of MF is mainly from the larger browning percentage of ENF.

In conclusion, after three years of drought from 2010 to 2012, there is increased browning of all woody vegetation types, and the browning area of ENF is larger than other woody vegetation types over Yunnan Province. The widespread loss of photosynthetic capacity of woody vegetation due to continuous seasonal droughts in Yunnan Province may represent a perturbation to the carbon cycle.

Acknowledgements.This work was funded by the National Basic Research Program of China (Grant No. 2012CB956202) and the Strategic Priority Research Program of the Chinese Academy of Sciences (Grant No. XDA05090200). The authors wish to thank Dr.YU Deyong and Dr. FANG Weihua for their help and suggestions.

Allen, C., A. K. Macalady, H. Chenchouni, et al., 2010: A global overview of drought and heat-induced tree mortality reveals emerging climate change risks for forests, Forest Ecol. Manag.,259(4), 660–684.

Barriopedro, D., C. M. Gouveia, R. M. Trigo, et al., 2012: The 2009-2010 drought in China: Possible causes and impacts on vegetation, J. Hydrometeorol., 13, 1251–1267.

Betts, R. A., Y. Malhi, and J. T. Roberts, 2008: The future of the Amazon: New perspectives from climate, ecosystem and social sciences, Phil. Trans. Roy. Soc. B, 363, 1729–1735.

Brando, P. M., S. J. Goetz, A. Baccini, et al., 2010: Seasonal and inter annual variability of climate and vegetation indices across the Amazon, Proc. Natl. Acad. Sci. USA, 107(33), 14685–14690.

Dai, A., 2011: Drought under global warming: A review, WIREs:Climate Change, 2, 45–65.

Davidson, E. A., A. C. de Araújo, P. Artaxo, et al., 2012: The Amazon basin in transition, Nature, 19(481), 321–328.

Herold, M., P. Mayaux, C. E. Woodcock, et al., 2008: Some challenges in global land cover mapping: An assessment of agreement and accuracy in existing 1 km datasets, Remote Sens. Environ., 112, 2538–2556.

Huete, A., K. Didan, T. Miura, et al., 2002: Overview of the radiometric and biophysical performance of the MODIS vegetation indices, Remote Sens. Environ., 83, 195–213.

Huete, A. R., K. Didan, Y. E. Shimabukuro, et al., 2006: Amazon rainforests green-up with sunlight in dry season, Geophys. Res.Lett., 33, L06405, doi:10.1029/2005GL025583.

IPCC, 2007: Contribution of working groups I, II and III to the fourth assessment report of the Intergovernmental Panel on Climate Change, in: Climate Change 2007, Pachauri et al. (Eds.),Cambridge University Press, New York, 104pp.

Justice, C. O., J. R. G. Townshend, E. F. Vermote, et al., 2002: An overview of MODIS land data processing and product status,Remote Sens. Environ., 83, 3–15.

Kileshye Onema, J.-M., and A. Taigbenu, 2009: NDVI-rainfall relationship in the Semliki watershed of the equatorial Nile, Phys.Chem. Earth, 34, 711–721.

Lewis, S. L., P. M. Brando, O. M. Phillips, et al., 2011: The 2010 Amazon drought, Science, 331, 554.

Malhi, Y., J. T. Roberts, R. A. Betts, et al., 2008: Climate change,deforestation, and the fate of the Amazon, Science, 319(5860),169–172.

Nepstad, D. C., I. M. Tohver, D. Ray, et al., 2007: Mortality of large trees and lianas following experimental drought in an Amazon forest, Ecology, 88(9), 2259–2269.

Phillips, O. L., L. E. Arag?o, S. L. Lewis, et al., 2009: Drought sensitivity of the Amazon rainforest, Science, 323(5919), 1344–1347.

Piao, S., P. Ciais, Y. Huang, et al., 2010: The impacts of climate change on water resources and agriculture in China, Nature,2(467), 43–51.

Saleska, S. R., K. Didan, A. R. Huete, et al., 2007: Amazon forests green-up during 2005 drought, Science, 318(5850), 612.

Samanta, A., S. Ganguly, H. Hashimoto, et al., 2010: Amazon forests did not green-up during the 2005 drought, Geophys. Res.Lett., 37, L05401, doi:10.1029/2009GL042154.

Wilhite, D. A., 1993: Drought assessment, management and planning: Theory and case studies, Biol. Plant., 36, 628.

Xu, L., A. Samanta, M. H. Costa, et al., 2011:Widespread decline in greenness of Amazonian vegetation due to the 2010 drought,Geophys. Res. Lett., 38, L07402, doi:10.1029/2011GL046824.

Yang, F., K. Ichii, M. A. White, et al., 2007: Developing a continental-scale measure of gross primary production by combining MODIS and AmeriFlux data through support vector machine approach, Remote Sens. Environ., 110(1), 109–122.

Zhang, L., J. F. Xiao, J. Li, et al., 2012: The 2010 spring drought reduced primary productivity in southwestern China, Environ.Res. Lett., 7, doi:10.1088/1748-9326/7/4/045706.