High Leachate Pollution at Indian Landfill Sites

 

by Hilary Haskell

India currently lacks an institutionalized municipal solid waste system and open landfill dumping proceeds without regulation, although an existing legal framework could be used to address this societal and environmental issue. A number of causes, both civilian and political, are at fault. Reliance on unregulated landfill dumping will likely continue, as this solid waste management practice tends to be the most cost-effective. However, landfills in India do not reflect the typical sanitary landfill seen in much of the developed world, lacking linings or covers that prevent groundwater pollution. Leachate liquid seepage generated by landfills due to rainwater or other infiltration pollutes groundwater. The extent of pollution depends on the permeability of landfills, distance to water table, and toxicity of the leachate. Bhalla et al. (2014) used data from the Ludhiana City, Punjab, municipal landfill site in India and the Leachate Pollution Index to determine Continue reading

Is Escaped Landfill Leachate Treatable?

by Hilary Haskell

Leachate containing Volatile Organic Compounds (VOCs) may be treated through a pump and treatment system at a landfill site. Following landfill closure, monitoring and extraction wells can be used to determine the effectiveness of this treatment method in preventing groundwater contamination. Martinez and Liu (2014) studied the Chestnut Avenue Landfill in Fresno, California, following its closure to determine if its pump and treat system had adequately reduced VOC concentrations in leachate. This Class III landfill had monitoring wells to collect groundwater data, but lacked a liner. Thus, the landfill provided an opportunity to research the behavior of leachate and its interactions with groundwater, and whether or not pump and treat systems could sufficiently remove VOCs from leachate. Continue reading

Drought Modeling Consistency and Discrepancies in Predicting Drought in the Future

Many models of soil moisture, drought indices, and precipitation-minus evaporation predict increased drought in the twenty-first century. Furthermore, precipitation, stream flow, and drought indices have demonstrated increased aridity since the 1950s over land. Still, there are major differences between observed data and model predictions. Sea surface temperature has been shown to influence land precipitation. However, coupled models have not reproduced recent regional precipitation changes in their predictions, which may be due to a lack of observed sea surface temperature data in these model stimulations. Dai (2012) demonstrates that the models reproduce the effects of El Niño-Southern Oscillation and the observed data on global mean aridity for 1923 through 2010. According to Dai, natural variations in tropical sea surface temperature not accounted for in the models cause the regional differences in observed and model-simulated aridity changes. Furthermore, he concludes that the observed global aridity changes through 2010 are consistent with model predictions, thus validating predictions of increasing severity and frequency of droughts over the next century. —Hilary Haskell
                  Dai, A. 2012. Increasing drought under global warming in observations and models. Nature Climate Change 3, 52–58.

                  Dai reconciles historical data and model projections of increasing aridity and drought in order to gain a more comprehensive understanding of global climate change’s effects on drought patterns. A variety of different drought indices quantify drought, yet yield different results, especially on smaller geographical scales. Dai uses precipitation, stream flow, and soil moisture fields to quantify meteorological, hydrologic, and agricultural drought, respectively. For historical soil-moisture data, the author used the Palmer Drought Severity Index (PDSI) Penman–Monteith (sc_PDSI_pm) equation to calculate the potential evapotranspiration of water out of the leaves of vegetation, given there is enough water for evapotranspiration to occur. The PDSI calculation is based on water-balance models of soil moisture forced with observed precipitation and temperature. This index is widely used in monitoring drought and paleoclimate reconstructions. The PDSI_PM is regarded as more accurate in its analysis than the PDSI_th calculation, because it includes temperature, precipitation, radiation, wind-speed, and humidity data in its calculation of potential evapotranspiration. Therefore, it provides a more comprehensive analysis of global warming scenarios. The sc_PDSI_pm calculation is used to determine the relative impacts of different factors (humidity, precipitation, etc.) affecting drought. By comparing the results of forced sc_PDSI_pm calculations that include or exclude various drought factors, each factor’s impact on drought can be determined. The author used two coupled climate model simulations based on future GHG emission scenarios: Coupled Model Intercomparison Project phase 3 (CMIP3), which was used in the Intergovernmental Panel on Climate Change Fourth Assessment Report and the new phase 5 (CMIP5).
                  For 1950–2010, Dai found that observed annual precipitation data and sc_PDSI_pm data had similar linear trends. Furthermore, these trends were similar to stream flow trends since 1948 for the world’s main river basins. The author noted some regional and quantitative differences between observed annual precipitation, sc_PDSI_pm, and stream flow data. However, these variations are still closely related. The similar linear trends indicate increased aridity over most of Africa, Southeast Asia, Eastern Australia, and Southern Europe, while there is increased wetness over the central U.S., Argentina, and northern high-latitude areas. This consistency in linear trends across independent measurements of precipitation and stream flow data indicate that these trends are accurate reflections of projected hydroclimatic variations in the future. Furthermore, this comparison also verifies that the sc_PDSI_pm is a reliable method for monitoring changes in aridity.
                  Warming trends since the 1980s have globally impacted the upward trend in drought areas, increasing these areas by about eight percent. This drying is attributable to warming patterns that cause increased evapotranspiration, especially over northern mid-to-high latitudes. However, increased aridity in Africa, Southeast Asia, Eastern Australia, and Southern Europe is mainly due to precipitation decreases. This decrease in precipitation is mainly caused by variations in sea surface temperature. Dai used The Hadley Centre Sea Ice and Sea Surface Temperature data set (SST) over varying periods of time to study the effects of (SST) on global drought. SST trends over the long-run are included in global warming predictions. However, over the course of a few decades, SST variations are absent in greenhouse gas (GHG) and aerosol-coupled model simulations. The absence of SST variations means that these natural variations are excluded from model predictions, albeit their timing and spatial patterns may be dependent on the initial conditions of the other factors included in the models. The lack of SST trends in current models makes the effects of SST changes irreproducible in models’ future predictions.
                  There are consistencies across soil moisture predictions for the coupled models and the sc_PDSI_pm calculations. Fourteen of the CMIP5 models analyzed in this study demonstrated decreases in soil moisture content in the top 10 cm layer of soil for most of the Americas, Europe, Southern Africa, most of the Middle East, Southeast Asia and Australia during the twenty-first century. The multi-model average predicts further decreases of 5 to 15% by 2080–2099. The sc_PDSI_PM using the same multi-model data reproduced the same increased aridity in soil moisture. However, the sc_PDSI_pm yields larger increases in wetness over central and eastern Asia and northern North America. In comparison to the CMIP5 model, the CMIP3 models, with some regional differences, also predicts the same increasing soil aridity for all seasons.
                  According to Dai, SSTs have large influences on land precipitation and drought. To demonstrate this influence, the author used maximum covariance analysis (MCA) of global SSTs at latitudes (40–60° N) and sc_PDSI_pm calculations at latitudes (60–75°N) based on observations and the CMIP models. The author’s goal was to reproduce the observed relationships between SST and sc_PDSI_pm by using MCA modes, in order to conclude whether the models could stimulate the recent drying trend. MCA analysis uses a standard singular value, obtained through decomposition of variables or inputs into a model from two separate fields. By decomposing this singular value, comparisons can be made between the two fields. For this study, Dai used MCA to compare difference in the SSTs and sc_PDSI_pm from observations and the CMIP models. Dai’s comparison of leading MCA modes excluded some of the unforced, irreproducible natural variation in the modes.
The first leading MCA modes (MCA1) based on observations and models represent global warming in relation to sea surface temperature. The temporal coefficient in this analysis is strongly correlated (r=0.97) between the observed global mean surface temperature and the SST MCA1 patterns, which are similar to the observed warming patterns over the oceans. MCA1 from the models show similar nonlinear global warming trends, with widespread warming over the oceans. The sc_PDSI-PM demonstrates short-term variability in these models. Considering the models, the observed mean global warming mode is evident in the GHG-forced CMIP simulations for both SST and sc_PDSI_PM, with a correlation of r=0.86 and a regression coefficient of .09566 between global means of sc_PDSI_pm anomalies from MCA1 observations. This finding suggests that inclusion of GHG emissions as a factor in global aridity modeling is a valid methodology based on its demonstrated consistency across global aridity change models and historical observations.
The second MCA modes (MCA2) calculated from both observations and models demonstrate similarity across spatial patterns. Both the observations and the models represent El Niño-Southern Oscillation (ENSO) through SST anomaly patterns. Furthermore, the time-based coefficient is highly correlated to the ENSO index. El Niño modes vary considerably across decades. Since 1999, ENSO observations for the central and eastern Pacific have become cooler than the time period from 1977 to 1998.
The spatial patterns for the sc_PDSI_pm were from the various much different than the observations, trend maps, and long term MCA1 (1950–2009). This discrepancy is attributable to large intermodal variations in the MCA1 from 1923–2010 caused by unforced natural variations and weak GHG-forced indicators of precipitation. These results indicate that individual models and observed global warming modes have large natural variation patterns unrelated to previous calculations that did not include GHG data. The MCA does not completely separate GHG-caused changes in precipitation and the changes in the sc_PDSI_pm from other natural variations included in the predictions. This finding is supported by the fact that until the year 2010, GHG -attributable variation was not very strong (4–6 % of total variance), compared with the natural variations forced in the models.
                  Dai suggests that the large regional differences between observations and individual model runs, the differences over West Africa, the USA, Brazil, Southern Africa, and eastern Australia, are the result of sample errors between actual realization of hydroclimatic patterns and natural variation not seen in the CMIP models. The major differences between the U.S. and Sahel stand out. The Sahel’s drying trend since 1950 is mainly due to decreased summer rainfall caused by warming in the South Atlantic Ocean relative to the North Atlantic, along with the warming of the Indian Ocean. This warming, coupled with dynamic vegetation feedback (not included in the CMIP models) leads to the discrepancy between the U.S. and the Sahel. CMIP3 models would predict the opposite warming pattern in the Atlantic Ocean and increased precipitation over the Sahel in the twenty-first century in the face of GHG caused global warming. CMIP5 models reproduce the decline in rainfall over the Sahel from 1950s–1980s. However, the CMIP5 models predict the decline with decreased magnitude and consider sulphate aerosols as the main cause of the decline. Other CMIP5 models do not include the effect of sulphate aerosol in the twentieth century. However, for the twenty-first century, the GHG effect will dominate over the aerosol parameter, and therefore, the drought over the Sahel may not actually take place.
                  Projected increased wetness over the U.S. is the result of the upward trend from 1950–1990 in precipitation levels. After the 1990s, however, the U.S. has become drier. This variation across decades can be attributed to Interdecadal Pacific Oscillation (IPO), with the warm phase of above normal SSTs in the tropical Pacific occurring around the year 1977, and the cold period taking place around 1999. IPO greatly influences US precipitation and drought, especially in the southwest. Anthropogenic data forcing does not coincide with these cycles due to the fact that these natural cycles are not included in predictions that depend on initial conditions of the coupled models. Therefore, these cycles are not reproducible in climate change predictions.
                  Differences between temporal and spatial patterns of the MCA1 mode for SST and sc_PDSI_pm in observed data versus CMIP5 data are largely attributable to model deficiencies in representing the effect of sulphate aerosols in the twentieth century, natural SST variations not taken into account by the CMIP models, and sampling errors among different GHG induced changes in the sc_PDSI_pm that are still relatively weak. MCA1 and MCA2 demonstrate the only statistically significant resemblance. This finding indicates that the global warming mode from the observations will likely become part of the GHG-induced warming model. The models are able to reflect the GHG-induced trend mode (MCA1) seen in observations  and the main ENSO mode (MCA2), therefore increasing confidence in the model predications of increasing drought patterns for the Americas Southern Europe, Southern and Central Africa, Australia, and Southeast Asia as GHG emissions contribute to continuing warming in the twenty first century. Still, these models are not reliably able to stimulate the precipitation and PDSI changes for these regions with valid certainty. The MCA1 patterns for sc_PDSI_pm for the twenty-first century are stable due to the large forced trend in the context of natural variations in temperature, precipitation, and other variables. The MCA1 patterns for sc_PDS_pm predict severe drought conditions by the late half of the twenty-first century, especially for densely populated areas such as Europe, the eastern U.S., Southeast Asia, and Brazil. If the model’s regional predictions are correct, the effect on these populations will be drastic.

Recovery of the Amazonian rainforest canopy after increasingly severe droughts

The Amazon rainforest has recently experienced increasingly severe droughts. Currently, there is little known about the long-term effects of these droughts that result in tree dieback, altered rainforest canopy structure, and increased rainforest flammability in the Amazonia.  Saatchi et al. (2013) use satellite microwave observations of rainfall from the Tropical Rainfall Measuring Mission and canopy backscatter data from the SeaWinds Scatterometer on board QuickSCAT to demonstrate that western Amazonia experienced severe water deficit during the dry season in 2005, and a subsequent disruption in canopy structure and decrease in canopy moisture. Furthermore, even with an increase in precipitation in the years following the drought in 2005, the decrease in canopy backscatter and thus alteration in the rainforest canopy characteristics and water content remained until the next major drought occurred in 2010. If droughts occur more frequently and severely in the Amazon rainforest due to climate change, this drought disturbance may cause considerable changes in the Amazonian rainforest canopy.—Hilary Haskell
                  Saatchi, S., Asefi-Najafabady, S., Malhi, Y., Aragao, L., Anderson, L., Myneni, R., Nemani, R. 2013. Persistent effects of a severe drought on Amazonian forest canopy. Proceedings of the National Academy of Sciences of the United States of America 110, 565570.

                  Saatchi et al. conducted this study to understand the extent and severity of the long-term impacts of droughts on the Amazonian rainforest. Currently, only short-term effects of drought events have been studied through ground and satellite data. The long-term impacts of drought on Amazonian vegetation have only been studied on controlled, small-scale (1-ha plot) field experiments. Recent studies of rainforest structure and density show an increase in tree die off and a decrease in above ground biomass following drought periods. Saatchi et al. were able to conclude that the Amazon rainforest’s response to severe droughts includes the increased mortality of large trees’ leaves and branches in the upper canopy, when soil water availability declined below a critical threshold. This response has subsequent impacts on the Amazonian rainforest canopy
                  Two severe droughts have occurred in the Amazon over the past ten years, as demonstrated by the anomaly in the Rio Negro River’s water levels. These severe droughts are typically associated with El NinoSouthern Oscillation (ENSO) events, which result in decreased soil moisture. This decrease in soil moisture endangers Amazonian vegetation by crossing critical water availability thresholds for long periods of time. If water stress continues, it causes higher tree mortality rates and increased flammability across the Amazon rainforest. The two major Amazon droughts considered in this study occurred in 2005 and 2010. The 2005 drought was not considered an ENSO drought, due to its temporal and spatial extent. The peak of this drought hit during the dry season and mostly impacted south-western Amazonia, unlike droughts caused by ENSO events. River levels during the drought in 2005 were the lowest to date. The tropical North Atlantic Ocean’s sea surface temperature increase could be a major cause of the 2005 drought. Anomalies in precipitation leading up to the droughts were also major contributing factors to the severity of the droughts studied. In Southern Amazonia, rainfall decreased by almost 3.2% per year in 19701998.The region also experienced negative precipitation levels during the decade leading up to the severe drought in 2005. This pattern of decreased precipitation could worsen in the future based on climate model predictions of the effects of climate change. Therefore, if Amazonian droughts continue to be as frequent and severe as those considered in this study, the overall ecosystem function and health in the Amazon rainforest will be affected.
                  Satellite spectral observations are used to detect changes in the rainforest’s upper canopy characteristics, such as greenness and leaf area, and may be used to determine long-term impacts of drought. Due to clouds and atmospheric aerosols, after the drought in 2005, the data from these observations was largely contradictory. Therefore, Saatchi et al. used data from two microwave satellite sensors to measure precipitation and canopy water content, in order to quantify the severity of the Amazonian droughts in the years 2005 and 2010, and their subsequent impacts on canopy water content and structure. The authors analyzed Amazonian precipitation with three Tropical Rainfall Measuring Mission indices that measure monthly precipitation (TRMM; 19982010). These indices include the dry-season precipitation anomaly (DPA), dry season water deficit anomaly (DWDA), and maximum climatological water deficit (MCWD). For the DPA, DWDA and MCWD, more negative values indicate more severe water deficit. The information from these indices is complementary and provides spatial-specific indicators of the extent and severity of deficit in the Amazon rainforest.
                  To measure the impact of water deficit on the Amazon rainforest, the authors used observations from the SeaWinds Scatterometer onboard QuickSCAT (QSCAT: 20002009), which uses a microwave frequency to provide backscatter measurements that demonstrate temporal and spatial variations of water content and structure in the rainforest canopy. Backscatter in this study was the reflection of the microwave frequency back to the QSCAT Scatterometer. The QSCAT signal at 2.1 cm wavelength and incidence angles of 50° penetrates 15 m into the rainforest canopy, and then scatters from leaves and branches of the upper canopy of trees. Backscatter measurements demonstrate biophysical properties of forests, such as water content and canopy structure. Changes over time (diurnal and seasonal) of canopy water content and seasonal leaf phenology of the vegetation’s lifecycle events have the greatest impact on the radar backscatter. Large-scale rainforest degradation and deforestation result in fewer trees within the rainforest canopy. When fewer trees are present in the rainforest, this change affects the canopy by creating gaps in its structure and altering its water content or biomass. These structural and biophysical changes are thus reflected in the signal of the backscatter. To represent the upper-canopy rainforest structure and water content, the authors used QSCAT backscatter data from dawn orbits, which monitor vegetation at its least-water-stressed time of day. The time series for this study were on normalized monthly and seasonally, and demonstrated spatial variations in upper-canopy-forest structure over the Amazon.
                  The authors experienced failure in November of 2009 with the QSCAT sensor scanner. The scanner stopped collecting data globally, which made it impossible for them to analyze changes in the canopy during the 2010 drought. QSCAT was reliable and without bias or sensor degradation prior to this failure. In order to still analyze the 2010 drought, the authors used TRMM precipitation radar backscatter data, which responds to surface moisture deep in the canopy, and scatters across soil and understory vegetation below the canopy through gaps within the rainforest.
                  Based on three indices from the TRMM data, Saatchi et al. found a severe drought over southwestern Amazonia in the year 2005. Of the approximate 5.5 M/km2 of forested area in the Amazon basin, about 1.7 M/km2 experienced DWDA less than −1.0 σ in 2005. About 0.27 M/ km2 experienced severe drought with DWDA less than −2.0 σ. In 2010, as the drought persisted, the spatial extent and severity increased, resulting in 2.6 M/km2 of the area subject to DWDA less −1.0 σ and 1.1 M/km2of forest area with DWDA less than −2.0 σ. The southwestern part of the Amazon typically experiences a dry season in normal years, but with this drought, the severity of the water deficit, with MCWD less than −300 mm, became larger. The wetter forests in the central part of the Amazon that had the most negative DPA and DWDA experienced low to moderate deficit with MCWD less than −100 mm in 2005 and 2010. The time span of these decreased precipitation anomalies was fairly short over the past ten years. However, the spatial extent in the southwestern Amazon suggested a pattern consistent with the low water levels measured in the Rio Negro River and other rivers within the southwestern Amazon basin.
                  The impacts of the 2005 drought on the forest canopy were captured with QSCAT backscatter time series data. The authors found a strong spatial correlation with water deficit anomalies from droughts observed by TRMM data for the same period and the QSCAT backscatter data, indicating that drought caused the change in rainforest canopy properties. The dry season in 2005 demonstrated widespread decline in rainforest canopy backscatter of 2.1 M/km2with anomaly of less than −1.0 σ, in the southwest Amazon, and an area of 0.77 M/km2  experiencing a decline in backscatter with anomaly less than −2.0 σ. The areas affected by the QSCAT anomaly included old growth rainforests to terra firme(rainforests not inundated by flood waters), from south to north respectively.
                  The authors compared the QSCAT and TRMM anomalies, and found significant correlations, with a 1−3 month lag between the decreased precipitation anomaly and subsequent canopy characteristic changes. The patterns of these correlations depended on rainfall, and varied over the Amazon. The lag ranged from zero to three months. The longest lags occurred over the southwestern region, due to the coupled effects of the naturally-occurring dry season precipitation variation in this Amazon region and the 2005 drought. The spatial variations from QSCAT and TRMM in 2005 with very negative QSCAT values are larger than similar areas with less severe WDA. To explain this discrepancy, the authors considered areas with maximum water deficit MCWD in 2005 that exceeded 300 mm and whether there was severe water deficit during the entire dry quarter with DWDA less than −3.0 σ. QSCAT backscatter reduction in the southwest region could be associated with the WDA developing in the dry season in 2005.
                  In order to analyze the impacts of the drought in 2005 over time, the authors used time series data of TRMM and QSCAT anomalies for the drought-affected southwestern areas of Amazonia. They found that after the year 2005, even though the area recovered in total precipitation, WDA remained negative for the dry season in 2006 and 2007, before recovery from this water deficit started with an abnormally wet year in 2009 over the entire Amazon (except the northeastern region). The water deficit became most severe in the southwestern region in late 2009. Areas affected by water deficit in southwestern Amazonia had low values in QSCAT backscatter in 2005 through the end of November 2009. To analyze longer term data trends, the authors used autoregressive moving-average models, a statistical model used for predicting future patterns in time series, on the order of 5% of the data points. Saatchi et al. found that the anomalies in QSCAT over time imply that the 2005 drought caused a step change in the back-scatter properties of the canopy, and that there was little recovery in subsequent years following the drought. These step changes indicate abrupt changes in the mean level of a time series of data. The authors considered whether this response could be attributed to hypothetical changes in sensor performance such as backscatter signal and calibration, but found that the response was unique to certain regions in western Amazonia that had the greatest water deficit anomalies. The annual QSCAT spatial patterns for the dry season support the finding of long-term reduction in backscatter from 2005−2009, thus indicating a change in canopy structure with more gaps in the canopy.
                  The QSCAT anomaly was most negative at a value between −2 σ and 3.5 σ in 2005, and remained a negative anomaly compared to previous years. In the years before the 2005 drought, the most negative anomaly had been a value of 0 σ. (P 1.2 σ, and the most severe water deficit peaking in 2010 at −1.5 σ to 2.0 σ.
                  To quantify changes in QSCAT data in comparison to the TRMM water deficit data, the authors compared average monthly anomalies for the southwestern Amazonia, and calculated the difference between QSCAT backscatter anomaly and TRMM monthly WDA. Doing so indicated that the difference in anomalies was zero (slope of zero) before the drought in 2005, and had a negative slope after 2005. This further substantiates the finding that there is a lag in recovery of QSCAT anomaly relative to the TRMM WDA in southwestern Amazonia, meaning that the recovery of the rainforest canopy does not immediately respond to increased precipitation. The accumulation of negative anomalies during dry summer months resulted in the largest decline in QSCAT backscatter in September of 2005.
The authors were able to extrapolate their findings over the entire Amazon by mapping the spatial distribution of the pixels with negative anomalies (less than −1.0 σ) for both QSCAT and TRMM data. There were significantly (P < 0.01) negative slopes between the QSCAT anomaly and WDA after the 2005 drought. Larger negative slopes indicate areas with a longer time period between water deficit recovery and subsequent canopy recovery, which suggest that droughts have lasting effect on canopy characteristics. All of the regions studied had an abnormally high number of fires in 2005 and the years following, indicating that there was lower water content in the rainforest canopy and more flammable dead vegetation, arising from a lack of precipitation. The impacts of these fires and rainforest degradation over the years 2005−2009 did not affect the QSCAT negative anomaly. Still, over 35% of the fires between 2005 and 2009 took place in QSCAT pixels that demonstrated a strong negative anomaly of less than −1.0 σ. The relationship between fires and water deficit is further supported by the fact that more than 78% of the rainforest in 2010 occurred in pixels with large negative slopes less than −1.0 σ, indicating a very negative difference between QSCAT and monthly TRMM WDA data. The prevalence of fires and areas that recovered slowly from drought coincided with regions that had large seasonality of backscatter in QSCAT data, indicating that these events occurred in mostly transitional and seasonal rainforests in the Amazon.
From this study, Saatchi et al. concluded that the QSCAT anomalies in backscatter data for canopy structure demonstrate the effects of the 2005 drought on the Amazon rainforest, and that the QSCAT data patterns are consistent with areas that experienced the largest water deficit in the driest quarter. QSCAT backscatter variability is the result of changes in the top layer of the canopy, which are exposed to water-vapor pressure deficits, making them more sensitive to droughts. The anomalies in radar backscatter indicate a decrease in upper-canopy biomass, layering, water content, and change in canopy structure attributable to drought disturbance. The authors concluded that the severity of these declines in canopy characteristics and altered canopy structure in 2005 is significantly greater than natural cycles and normal seasonal declines in QSCAT backscatter resulting from phonological life cycles and canopy water content. The authors also concluded that because the QSCAT negative anomaly persisted over most of the western Amazon after the 2005 drought that droughts have lasting effects on the rainforest canopy. Due to the magnitude of the drought disturbance, there was a lag in recovery of the rainforest canopy in terms of biomass and roughness (canopy layering) after rainfall returned to normal levels.
Saatchi et al. discussed the need for more extensive surveys and airborne observations to better assess the decline of backscatter and its relation to rainforest disturbances evident in canopy biomass and layering. During the most severe periods of drought, tree leaves should wilt and shed, therefore causing a decline in net primary production of the rainforest canopy. Within a year of a drought event, recovery of the canopy to its original pre-drought status should occur. In central Amazonia, this recovery did occur where QSCAT backscatter anomalies recovered quickly after water deficits post-drought in 2005. However, the southwestern Amazon rainforest demonstrates a lasting decline in canopy structure, and a recovery time of three to four years, as indicated by delays in recovery of QSCAT backscatter data. Canopy structure in these instances is affected by branch dieback and tree fall creating holes in the canopy structure.
In order to verify the findings of this study, the authors attempted to find in situ observations over the region affected by the drought in 2005. From previous studies, there is consensus that large or emergent trees have higher mortality than small trees. The amount of light penetrating into the understory below the rainforest canopy is affected by the gaps in the canopy that occur from tree fall. After precipitation recovery occurs, tree fall from drought events can increase light availability, and therefore promote understory vegetation and pioneer species productivity. The authors found that in situmeasurements indicate that large trees die off more rapidly for a few years after drought. This finding suggests that canopy structure and biomass decline before recovering gradually with the emergence of more trees. However, this recovery process is slow, and may require a considerable amount of time to return to the pre-drought state. From their findings in this study, the authors suggest that the western Amazonia experienced a large-scale canopy disturbance due to the 2005 drought, resulting in the decline of and canopy structure and biomass that continued with slow recovery for the next few years. In the future, the authors suggest further research with direct examination of the effects of severe drought on forest plots in western Amazonia, to verify their conclusions.
Due to rainforest die-off from large-scale drought disturbance, there may be an impact on the rainforest’s carbon dioxide absorption and release processes. This impact is caused by the decay of wood from increased tree mortality. During the years 1995 through 2005, a decline in plant water availability created water stress before the onset of the 2005 drought, therefore playing a role in the canopy disturbance after the 2005 drought occurred. Because of the drought in 2005, and a local drought in 2007, much of the soil in the southwestern Amazonia may not have reached water content capacity that would encourage canopy recovery. Variability in the natural dry season over the southwestern Amazonia since the late 1980s and early 1990s may also have been contributing factors. Sea surface temperature is a major contributor to the negative anomalies and year-to-year variation.
                  Saatchi et al. concluded their study in 2009. From their findings, they suggest that the Amazon rainforest canopy response from the drought in 2005 was repeated in 2010. For the year 2010, there were no QSCAT data available. Therefore, further drought disturbances may have affected the rainforest canopy that was yet to recover from previous drought and decreased water availability. TRMM-PR backscatter anomalies indicate that surface moisture in southwestern Amazonia dropped significantly in 2010, and lasted longer than the dry season, further stressing the canopy that had not yet fully recovered. If this pattern of drought continues to occur on a 5–10 year time scale, or became even more frequent, the Amazonian rainforest canopy will be exposed to drought consistently, and the rainforest canopy will be slow to recover in structure and function. Southwestern Amazonia has been particularly subject to severe effects of rainfall variability during the last ten years, thus indicating that this region is perhaps the most vulnerable to large scale rainforest degradation based on climate change.

The Effects of Soil Moisture on Precipitation

During the day, vegetation and soil moisture affect the absorption of the sun’s heat between soil and the atmosphere. When droughts occur, soil lacks enough water for evapotranspiration to occur at its full potential. With decreased evapotranspiration, the lower atmosphere becomes hotter and drier. Therefore, soil moisture is an important factor in the development of storms that return moisture to soil through feedback loops. However, there is uncertainty regarding how soil moisture affects storm formation globally, due to a lack of observed evidence and uncertainty in large-scale models. Taylor et al. conducted a worldwide observational analysis of the relationship between soil moisture and precipitation (2012). The authors were able to conclude that afternoon rain falls predominantly over soils that are relatively dry compared to the soil in the surrounding area over certain spatial scales and seasons. This finding is most obvious in semi-arid regions where fluxes in heat and humidity are most sensitive to soil moisture and there is frequent convection of heat and moisture between the soil and atmosphere. The authors found that their results indicate increased afternoon moisture caused by increased heat flux into the atmosphere over drier soils, as well as possible variability in soil moisture over a scale of 50-100 km. This study suggests no positive feedback loops of increased precipitation coupled with wetter soils over the spatial scale studied. In contrast, six state-of-the-art global weather and climate models predict positive feedback between soil moisture and precipitation. The authors largely attribute this discrepancy to excessive drought predictions in large-scale hydroclimatic models. —Hilary Haskell
Taylor, C., de Jeau, R., Guichard, F., Harris, P., Dorigo, W., 2012. Afternoon rain more likely over drier soils. Nature 489, 423-426.

Taylor et al. evaluated the response of daytime moisture convection between soil and the atmosphere to soil moisture anomalies. The authors used global observational data sets of both surface soil moisture and precipitation at a resolution of 0.25° x 0.25° on daily and 3-hourly time scales in order to analyze the location of afternoon rain events relative to soil moisture, before the rain events occurred. Soil moisture data was retrieved between 60°south and 60° north from the Advanced Microwave Scanning Radiometer for EOS and the MetOP Advanced Scatterometer.  The authors considered whether rain is more likely over soil that is either wetter or drier than other soil in the surrounding area. This methodology was also used in six global models of climate projection.
Soil moisture affects precipitation over a variety of time periods and geographical areas. When drought occurs, there is less evapotranspiration of water out of plant leaves into the atmosphere, causing overall atmospheric moisture content to decrease. Less atmospheric moisture leads to decreased precipitation. Anomalies in soil moisture disrupt atmospheric heating patterns, and thus synoptic-scale (large scale, usually 100 km or longer) atmospheric circulations and horizontal water vapor transport from the oceans. Small scale disruptions, such as regional precipitation and the formation of convective clouds due to atmospheric instability in heat and moisture content, can be affected by fluctuations in these factors throughout the day. Both convective clouds and rainfall occur due to the sun’s uneven heating of the earth’s surface and atmosphere, thus leading soil moisture to evaporate and turn to water vapor. When the water vapor rises and cools, it condenses back into liquid form in clouds, before eventually becoming heavy and falling in the form of precipitation. In an undisturbed atmosphere, surface feedback between soil and the atmosphere is determined by atmospheric temperature and humidity. Mesoscale (10-100vkm in length) variability in soil moisture can create feedback loops through daytime circulation of soil and atmospheric moisture.
            A variety of studies have indicated different correlations between soil moisture and precipitation. In Illinois and West Africa, studies found positive correlations, demonstrating a positive feedback between soil moisture and precipitation. Another study found increasing frequency of heavy convective rainfall coupled with high rates of evapotranspiration in North America. Sensible heat fluxes are the fluctuations of heat between the Earth’s surface and the atmosphere through conduction and convection at the boundary between the atmosphere and the soil. Satellite cloud data indicate an increase in afternoon precipitation frequency over areas with increased sensible heat fluxes from mesoscale circulations, caused by soil moisture or vegetation cover.
            Regionally, climate models consistently predict feedback between soil moisture and rainfall. Soil moisture limits evapotranspiration when convection occurs, due to the lack of available water in the soil for both processes. However, the strength of these feedback loops is less consistent, indicating that there is uncertainty in surface flux sensitivity to soil moisture and the impact of surface fluxes on the convection in the boundary layer between the atmosphere and soil. The authors found that whether the feedback relationship between soil moisture and precipitation is positive or negative depends on the model’s spatial resolution. This resolution is impacted by convective parameterization, which provides a quantitative measurement of atmospheric instability, a triggering mechanism for convection, the net effect of latent heating on the local atmosphere, and the amount of water vapor available in a convective process. This study suggests that smaller spatial scales produce different predictions of soil moisture and precipitation feedback patterns due to erroneous climate models and domination of large-scale atmospheric conditions.
            The authors focused on afternoon precipitation development, when sensitivity of convection of moisture and heat to land conditions is maximized. A precipitation event was defined as a 0.25 x 0.25 pixel location within a box measuring 5 x 5 pixels, and precipitation that exceeded 3 mm. To ensure that soil moisture measurement preceded rainfall, the authors excluded pixels with more than 1 mm of rain in the hours preceding the precipitation event. The authors also excluded locations with topographic heights above 300 m and regions containing water bodies or strong soil moisture gradients. Therefore, mountainous and coastal areas were not included due to their effects on mesoscale precipitation, or tropical forests due to the inability to retrieve soil moisture data beneath dense vegetation. Each maximum afternoon rainfall (L max) was paired with one or more pixels in the box where afternoon rainfall was at a minimum (L min). The authors then measured the difference in pre-rain-event soil moisture ∆Se by subtracting L min from L max, with an adjustment for the climatological mean soil moisture from both the control and experimental locations. The strength of the soil moisture’s effect on precipitation was quantified by a sample of precipitation events. The events were analyzed for predictability by comparing how unexpected the observed sample mean value of ∆Se was relative to a control sample, Sc, from the same location on non-rain-event days. The control sample was constructed from daily soil moisture differences between L max and L min, using data from the same month but non-rain-event areas. This was expressed as a percentile of typical values δe. Rain over drier soils is indicated by values less than ten, while rain over wetter soils is indicated by values greater than ten.
            The authors concluded that globally, 28.9% of the areas analyzed had percentile values less than ten (indicating rain over drier soils), compared to an expected frequency (which assumes no feedback between soil moisture and precipitation) of 10%. Lower percentiles are in semi-arid to arid regions, such as North Africa, Eastern Australia, Central Asia, and Southern Africa. This finding indicates that rain over drier soils is a common phenomenon. When computing the mean difference in soil moisture before rain events for both the experimental and control groups, this same outcome occurred with 99% statistical significance across all continents and their climate zones. However, upon repeating the analysis after lowering the spatial resolution from 0.25 ° to 1.0° degrees, this produced only about one-tenth of the number of events compared to the 0.25° data.  Still, for this repeated analysis, the authors found that afternoon rain over drier soils was more likely, and that this finding was statistically significant for parts of North Africa and Australia.
            According to Taylor et al., the satellite-driven data sets are subject to error in providing data for the rain events. However, analyzing data from many precipitation events should still yield accurate estimates of δe. Additionally, satellites are a valuable data source in that they are able to demonstrate the spatial structure of rainfall. This further substantiates the authors’ methodology by providing consistency in the spatial variability of soil moisture and rainfall studied in the independent data sets.
            After concluding that rain tends to fall over the drier soils within regions, Taylor et al. also studied whether drier soils are consistent with land surface feedback between soil and atmospheric moisture. For this to be the case, evapotranspiration must be limited by soil water deficit. The consistency in land surface feedback and drier soil is only present during some seasons and in regions where water stress coincides with the convective activity. Low percentiles of  δe occur in relatively dry areas, and are due largely to convective storms. The most negative values occurred in the driest mean conditions. This finding is consistent with soil moisture feedback, meaning that the sensitivity of heat fluxes (sensible and latent) to soil moisture increase as mean soil moisture decreases.
            For there to be a soil moisture feedback, there must be strong daily tendency for afternoon precipitation to fall over soils drier than other soils in the surrounding region. The authors repeated their analysis a third time and evaluated the onset of precipitation at varying times after soil moisture observations were taken at 1:30 P.M. local time. This reanalysis produced values of δe   that demonstrate a daily cycle, with the most negative values occurring during daytime (between 12:00 P.M. and 3:00P.M.) over dry soils and positive values during nighttime (between 9:00 P.M. and 3:00 A.M.) over wetter soils. The early afternoon minimum over drier soil is consistent with negative soil moisture feedback. This type of feedback occurs when convective instability, effects of surface properties on the planetary boundary layer between the soil and the atmosphere, and mesoscale flows are maximized. However, the mechanisms to explain the positive values at night time are less obvious. After dusk, day-time surface-induced flows and thermals, columns of rising 
air in the lower altitudes of the atmosphere created by the uneven heating of the Earth’s surface by the sun, do not persist. However, nocturnal humidity anomalies do persist and last longer depending on the spatial scale of surface features and wind conditions. At night, pre-existing, fast-moving convective systems are the most influential factors in producing positive night values
            The authors repeated their analysis a final time using 3-hourly diagnostics from six global models, ranging in resolution from 0.5° to 2.0°. In contrast to the previous analyses, this analysis found a strong preference for rain over wet soils for most parts of the world. Only one of the models produced more than the expected 10% results of with P < 10 (with more negative
δe valuesindicating more rainfall over drier soil). The P < 10 areas were also not distributed as globally as they were in the authors’ previous analyses, and were concentrated in the mid-latitudinal region. This analysis across the six global models produced the opposite result compared to the author’s previous analyses. Increased precipitation over wetter soils, especially in the tropics, indicates the failing of convective parameterizations to represent land feedbacks on daytime precipitation. This failure is largely attributable to the lag in the daily cycle of precipitation, with rainfall starting several hours too early. Furthermore, this failure is amplified by a lack of clouds within the boundary layer between soil and atmosphere in some models. Convective precipitation occurs rapidly and intensively over a limited area, due to unstable moisture and heat in the atmospheres, and is highly responsive to the daytime increase of moisture instability. This over sensitivity causes precipitation to occur faster over wetter soils, indicating a positive feedback cycle. When rainfall occurs prematurely in daily cycles, there are other daytime factors that contribute to convection of moisture into the atmosphere are limited in their effect.
            Still, Taylor et al.’s finding that afternoon rain over locally dry soils on scales of 50 -100 km is consistent with studies using remotely sensed data. The authors’ failure to find instances of positive feedback is an indication that surface-induced flows over 50-100 km are important in triggering convection. In addition, the authors cite mixing processes in the formation of convective clouds as an important contributing factor. However, neither of these factors is currently included in the one-dimensional analyses utilized in this study. The authors also consider whether models that use convective parameterizations represent the aforementioned processes adequately. The models analyzed in this study (HadGEM2, CNRM-CM5 and INMCM4) do not resolve soil moisture structures across 50-100 km or their impacts on triggering the convective heat and moisture flux between the atmosphere and soil. Still, all the models support the finding that rain tends to fall over wetter soils, despite the lack of observational data. The authors do not aim to conclude that soil moisture feedback is negative for temporal or spatial scales not analyzed in the study, such as those on large spatial scales. Taylor et al. postulate that the accumulation of moisture in the lower atmosphere from transpiration over land surfaces may provide more favorable dawn conditions for daytime convection of moisture into the atmosphere than the equivalent accumulation in a region affected by drought. Or, the large-scale response to soil moisture may dominate regional responses in some areas. The authors conclude that the sensitivity of convection patterns used in the six climate models used in this study is inaccurate, and contributes to the tendency of large-scale models to erroneously extend measures of drought periods and exaggerate soil moisture feedbacks in climate systems.

Physiological Mechanisms of Delayed Drought-Induced Forest Mortality

Forest mortality could become more severe and frequent due to global change in hydroclimatic patterns. Tree deaths will affect the global carbon cycle, ecosystem services, and levels of biodiversity. However, the mechanisms by which trees die-off over the course of the time are largely unknown. Anderegg et al. examine the physiological basis for the delayed, widespread die-off of trembling aspen (Populus tremuloides) across North America post-drought (2012). The authors studied experimental and observed data for both dying and healthy trees to assess physiological performance and accumulated hydraulic damage (hydraulic deterioration) demonstrated by xylem water conductivity. Cavitation fatigue caused hydraulic damage to persist or increase in dying trees, and was found to predict the probability of tree mortality over the course of the years studied. The authors also conclude that surviving regions of forests that have already been subjected to drought are the most vulnerable to future drought. Increased drought vulnerability affects ecosystem stability, biodiversity, and ecosystem carbon balance. Drought stress accumulation and repair affects tree mortality, which has subsequent implications for forest ecosystems.—Hilary Haskell
Anderegg, W., Plavcova, L., Anderegg, L., Hacke, U., Berry, J., Field, C. 2012. Drought’s legacy: multiyear hydraulic deterioration underlies widespread aspen forest die-off and portends increased future risk. Global Change Biology 19, 1188-1196.
Anderegg et al. studied the physiological basis of delayed widespread die-off of trembling aspen (Populus tremuloides) post-drought. The authors studied trees in Colorado’s San Juan national Forest, which experienced some of the worst sudden aspen decline (SAD) of any area in western Colorado. The forest studied has a mean annual temperature of f 3.2 °
C and mean annual precipitation 508 mm. In the forest, precipitation falls as either snow (November through May) or monsoon summer rain (July through September). Seasonal droughts typically occur in early summer, and represent peak water stress for the forest. The study’s specific site consisted of seven aspen clone groups with mortality rates ranging from healthy to SAD trees within 100 m of one another. The authors measured crown dieback, the percent of a tree’s branch tips within the crown of the tree (above ground parts: leaves, branches, trunk, etc.) that are dead as an indicator of tree health. Two categories of ramets, an individual specimen of a clone, were compared in this study: SAD affected (ramets with less than 50% crown dieback) and healthy (ramets with less than 20% crown dieback). Tree death in this study was designated as 100% crown dieback. The hydraulic properties of the trees studied are largely dependent on xylem, the woody vascular transport system that transports water from the soil throughout the plant.
The authors used observed and experimental evidence of trees’ accumulated physiological hydraulic changes to assess whether hydraulic properties can be used to predict inter-year stem mortality. Stem mortality is an indicator of subsequent tree death. These changes in tree’s hydraulic properties could thus explain delayed and multi-year aspen die-off after exposure to drought.  Anderegg et al. sought to explain whether different levels of water stress, accumulated growth of xylem, vulnerability of xylem vessels to cavitation or some combination of these factors caused the differences between the accumulated hydraulic properties of trees that survived and those that died.
The authors measured crown dieback and hydraulic conductance for thirty-six ramets, both SAD and healthy, between August 2010 and July 2011 in order to assess whether the hydraulic performance demonstrated by xylem water conductance affects inter-year tree mortality probability. Logistic regression techniques were used to estimate the probability of a ramet dying during this time period, compared to baseline native xylem water conductance in tree branches. During August 13–14, 2010, between 12:00 P.M and 2:00 P.M, the authors collected branch networks (branches, petioles, and leaves) from thirty-six clones that were farther than 10 cm from the mid canopy of the forest. After collecting the branches, they were kept moist by water in dark plastic bags. Then, the branches were re-cut under water. Using the vacuum method, the authors calculated hydraulic conductance. These hydraulic conductance measurements reveal instances of hydraulic failure where the xylem fails to transport water along the entire branch sample.
Anderegg et al, studied seasonal levels of plant water stress using xylem tension measures, between the dates of June 17–19, July 16–18, and August 17–19. These periods reflect the annual peak water stress period in June and July, and end of the peak period in August. Weather was generally clear and sunny with temperatures ranging from 16 and 19 °C when taking the samples. For each set of clones, two randomly selected branches per ramet, and three ramets per class (healthy versus dying) were evaluated. The same method was used to maintain moisture for sample branches as used in measuring inter-year tree mortality. Five minutes after collection, the authors removed twigs longer than 30 cm, two branch forks away from the branch’s break from the trunk. They used this twig to measure branch xylem tension, using a Schlander-type pressure chamber. The process was repeated three times per branch for consistency and accuracy. Branch xylem tension was taken at two points during the day: dawn (when the tension is lowest and plant water is equal with soil water, absent nightly transpiration of water out of the tree’s leaves), and at midday (when highest daily tensions occur, due to transportation).
Another parameter used to predict the probability of tree mortality over time was changes in growth, which also partially accounts for differences in conductivity. The authors measured branch growth and xylem vessel diameter for both the healthy and SAD ramets. Branch growth was measured using annual growth ring width data from the years 1998–2011. All branches studied were eighteen to thirty years old and fully developed before the drought hit from 2000–2003. Using a sliding microtome, 40<!–[if gte msEquation 12]>μ<![endif]–> branch cross-sections were cut and observed under a light microscope (DM3000; Leica) in order to measure ring width. Nine branches per treatment from a June 2011 sample were analyzed for both healthy and SAD treatments. The vessel diameters apparent in growth rings from the years 2003 through 2010 were observed. 350 vessels were measured per growth ring for healthy and SAD treatments. This finding suggests that lasting hydraulic changes are not apparent in vessel diameter. Furthermore, branch growth demonstrated resilience after three years of decreased growth post-drought. Both hydraulic changes and growth are affected by carbon balance within a tree. Growth rates parallel carbon uptake through a positive relationship. This study suggests that changes in hydraulic physiological processes lag behind droughts, but are reversible and play a minor role in hydraulic performance for this species.
Xylem’s vulnerability to cavitation over the course of multiple years limits the effectiveness of cavitation repair as well as restoration of hydraulic conductivity. Cavitation occurs within the xylem when water tension becomes so extreme that the water vaporizes, and the resulting air bubbles fill the water-conducting vessels, thus blocking hydraulic transport. To examine changes in xylem vulnerability, the authors used the standard vulnerability curve technique that measures the percent loss conductivity (PLC) of a branch sample as a function of artificial xylem tension. To determine the vulnerability curves, the authors used either centrifuge or air injection methods to generate simulated cavitation-inducing pressure differences in the tree branches. Embolisms (air blockages) were flushed out of the xylem via vacuum infiltration prior to measuring for the vulnerability curves. The authors used the maximum conductivities for both native and flushed specimens to ensure that flushing out embolisms did not bias results. The lab-tested xylem vulnerability curves were plotted against the baseline native hydraulic conductivity before and after vacuum infiltration, in order to validate the findings. Increased vulnerability to cavitation after exposure to drought stress is called “cavitation fatigue,” indicating a physiological threshold in a tree’s ability to cope with repeated periods of drought. In vulnerability curve analysis, cavitation fatigue is evident in conductivity changes. In the years after experimental drought, increased PLC occurred for ramets subjected to drought, despite similar xylem water tensions and water availability.
                  Finally, the authors monitored xylem water conductance for trees recovering from water stress induced by an experimental drought. To do so, the authors followed the hydraulic performance and health of mature individual ramets during and after the experimental drought. To simulate drought, two separate plots of mature aspen clones were excluded from rainfall from June to August, 2010. Therefore, the mature ramets for each treatment in the experimental group were subject to water stress over the course of the growing season. Anderegg et al. measured xylem tensions and hydraulic conductivity for the month of July in 2010, 2011, and 2012 under these simulated drought conditions. Their findings indicate that delays in tree mortality are associated with a shift toward increased vulnerability to cavitation in ramets exposed to drought. This cavitation fatigue is important in understanding why delayed hydraulic changes are related to forest die-off.
                  Anderegg et al.used ANOVA statistical analysis for time-series data and repeated measures of xylem tensions, loss of conductivity, and vulnerability. Direct group comparisons utilized Student’s t-stest, given normality from the Shapiro Wilkes test.
Twelve of the thirty-six ramets died between August 2010 and July 2011, including both healthy and dying trees. This finding suggests that differences in hydraulic conductivity can be linked to probability of mortality, even years after drought stress. With this finding, the authors were able to create a logistic regression based on August 2010 native branch xlyem water conductance versus the probability of ramet mortality between August 2010 and July 2011. Anderegg et al. concluded that native hydraulic conductance of branches could predict inter-year mortality between 2010 and 2011. The authors tried to determine what caused these differences in conductance. Initially, they considered water stress in dying trees, further amplified by root mortality. However, they found that SAD and healthy ramets did not have significantly different xylem tensions (P= 0.47), meaning that dying and healthy trees did not have different levels of water stress during the typical seasonal periods of early summer months.
To determine the influence of new xylem growth on the hydraulic disparities between dying and healthy trees, the authors tested differences in xylem diameter and growth to find that xylem vessel diameter was smaller overall in the year 2003 than in 2010 for both SAD and healthy branches. Branch growth decreased significantly in SAD ramets for three years after peak drought severity (P=0.01), but branch growth recovered by 2007, and remained constant through the year 2010. This finding suggests that growth in trees can be resilient post-drought, given adequate water supply and decreased hydraulic stress
                  Dying trees’ xylem did not grow slower or experience higher water stress. Still, the authors attempted to determine whether water stress was more damaging for dying than healthy trees. Xylem’s vulnerability to embolism was significantly higher for SAD dying branches than healthy branches (P=0.001). The authors also found this to be true for differences in embolism vulnerability apparent in hydraulic conductivity. Xylem pressure became fifty percent less conductive between pressure ranges of –1.0 MPA for SAD ramets and –2.3 MPA for healthy ramets. SAD branches had substantially lower water conductivity overall. After exposure to drought and thus higher xylem tensions, healthy aspen clones’ vulnerability was higher than that of health clones not exposed to drought with lower xylem tensions.
                  The experimentally-induced drought produced results similar to those seen in SAD areas. Tree mortality lagged post drought, and trees experienced many of the same physiological hydraulic changes. The time it took for crown dieback to occur lagged in the experimental study, as it did in SAD areas. In 2012, two years after water exclusion and drought ceased, moderate differences in dieback between control and drought-exposed ramets were significant (P = 0.1). After more than two years, xylem tensions for both control and drought-subjected ramets were almost identical, and only varied with seasonal rainfall. PLC during experimental drought conditions increased for drought ramets, and remained elevated after the experimental drought stress ended. In 2011 drought-stressed ramets had significantly higher PLC than control ramets (P = 0.01). Therefore, this study suggests that despite identical initial xylem tension conditions from the 2010 experiment, when ramets are subjected to drought, they demonstrate increased vulnerability to cavitation and subsequent mortality post-drought
Anderegg et al. found that delayed and accumulated physiological hydraulic changes affect wide-spread, drought-induced aspen die-off. The hydraulic performance of a tree, demonstrated by xylem water conductance, is an important predictor of tree mortality over time. Dying ramets do not experience different levels of water stress during seasonal drought, indicated by the lack of seasonal xylem tension changes. Furthermore, xylem tensions were similar for drought-exposed and control trees for two years following the experimental drought study. The authors suggest further study into the physiological mechanism behind dying trees’ similar xylem tensions, whether this phenomenon be due to stomatal regulation or concurrent root and leaf mortality. Root mortality does occur with SAD, which may be important for feedback effects from water stress. However, the balanced timing of root and crown mortality requires further research.
This study discusses the role of accumulated physiological hydraulic changes in xylem water conductivity (hydraulic deterioration) in tree die-off.  This deterioration is not reversible, even if normal hydroclimatic conditions are restored. Hydraulic deterioration is related to SAD tree mortality, due to trees’ vulnerability to seasonal water stress prior to severe drought. Trees can tolerate these seasonal droughts, but the accumulated damage from these annual droughts is too much to overcome in the face of severe drought. Logistic regression can be used to predict the probability of mortality across years, based on accumulated hydraulic deterioration.
The authors discuss other possible reasons for die-off, including branch scarring and elevated insect attack that exacerbate hydraulic deterioration. In addition, SAD areas had increased levels of fungal pathogens that could affect hydraulic deterioration. Still, multi-year drought feedback effects, repair of damage, and accumulated stress are crucial to understanding forest’s vulnerability to drought. The ability to repair drought-induced damage is just as important to a tree’s survival as the severity of the damage itself. These results from trembling aspen can be extrapolated to other delayed, post-drought tree death incidences in forest ranging from Canada to the tropical Amazon. The physiological mechanisms at work in trembling aspens are widely applicable across other species. 

The Relationship between Embolism Resistance and Droughts for Global Forest Ecosystems

Global climate change predictions indicate increasing temperatures and shifts in rainfall patterns, resulting in increased severity and frequency of drought. These droughts are likely to cause forest decline around the world, resulting in decreased net primary productivity and plant mortality, largely attributable to hydraulic failure. Through a process called embolism, drought stress causes gas to become entrapped in the vascular systems of plants. Embolism thus prevents water transport necessary for photosynthesis to occur in plants. If a plant is unable to conduct photosynthesis, it will eventually desiccate and die. The threshold limits for hydraulic failure across different species and environments is largely unknown. Choat et al. (2012) compared the vulnerability of a variety of woody species based on drought-induced embolism. The authors found that seventy percent of the 226 forest species studied operate within narrow hydraulic safety margins against damaging levels of drought stress. If these margins are exceeded, long term implications including decreased productivity and survival are affected. These safety margins are largely independent of mean annual precipitation, thus demonstrating that there is global convergence in the vulnerability of forests to drought, meaning that all forest biomes are equally susceptible to hydraulic failure, regardless of initial rainfall environments.
—Hilary Haskell
            Choat, B., Jansen S., Brodribb, T., Cochard, H., Delozn, S., Bhaskar, R., Bucci, S., Feild, T., Gleason, S., Hacke, U., Jacobsen, A., Lens, F., Maherali, H., Martinez-Vilalta, J., Mayr, S., Mencuccini, M., Mitchell, P., Nardini, A., Pitterman, J., Brandon Pratt, R., Perry, J., Westoboy, M., Wright, I.,

Zanne, A., 2012. Global Convergence in the Vulnerability of Forests to Drought. Nature 491, 752–55.

Choat et al. quantified comparisons of plants’ sensitivity to drought stress using the strength of the liquid (hydraulic) connection between soil and leaves through water-transporting tissue within the vascular xylem tissue system of plants. Plant leaves’ stomata act as gateways that conduct the exchange of gases such as water vapor, carbon dioxide, and oxygen from within the plant to the atmosphere. When soil becomes arid, stomata regulate water loss from the leaves in order to maintain the xylem pressure within a range that will prevent embolism.  After prolonged periods of drought and drying soil, stomatal closure slows but does not halt the decrease in xylem pressure and hydraulic capacity. If soil water is not replenished, complete hydraulic failure occurs, causing the plant to desiccate and die.
Embolism is essentially air blockages within the xylem of plants. The process by which embolism occurs is dependent on the xylem structure of woody plants. Cavation occurs when liquid water changes phases to become vapor. This process occurs because water in the xylem is under negative pressure. The air emboli (blockages within the xylem) prevent or reduce the plant’s ability to transport water from soil to sites of photosynthesis necessary for the plant to survive. Woody plants that are able to survive and recover after sustained drought are generally considered embolism resistant. Embolism resistance varies across species, and largely depends on xylem structure. The relationship between xylem pressure and loss of hydraulic conductivity due to gas emboli defines the plant’s resistance to embolism. The index used to quantify this relationship is water potential measured in megapascals (MPa) (Ψx), at which fifty percent loss of conductivity within the xylem structure occurs. A slight drop in Ψx at this point will confer a substantial reduction in hydraulic function. When Ψxfalls below Ψ50, the water transport of the xylem is impaired and the plant is at risk of embolism. After embolism occurs, there are irreversible reductions in productivity, tissue damage, and death. Embolism resistance is also described by a vulnerability curve, which shows the percentage loss of hydraulic conductivity (PLC, %) as a function of decreasing xylem pressure measured by ΨxMPa. Decreasing xylem pressure reflects increased drought stress, and therefore aridity.
Xylem structure can acclimate to environmental variation while it is still developing; however, subsequent adaptation is impossible, due to the fact that xylem tissues are dead at maturity. For plants subject to drought or increased aridity after maturation, this poses a serious threat to the organism’s survival. Therefore, embolism resistance is important in determining the limits of drought tolerance for woody species, and thus, predicting drought-induced forest decline globally and regionally. Forests in this study were defined as Mediterranean, savanna, and woodland environments. Therefore, the plant species analyzed included trees, shrubs, and lianas.
The authors used a database of 480 woody species with Ψ50 to compare forest biome vulnerability to drought-induced hydraulic failure. The species studied came from a variety of climates, with mean annual precipitation (MAP) ranging from 300 to 4,500 mm and mean annual temperature from −4 to 27°C. Climate data for the study was taken from the WorldClim database of CRU climate data base. 384 Angiosperms (flowering plants) and 96 gymnosperms (mostly conifers) were taken into account separately due to structural differences
between the two types of plants. Shoat et al. found a significant (PΨx (Ψmin) for plants under natural conditions and Ψ50 for the angiosperms and gymnosperms studied. This finding suggests that there is a relationship between embolism resistance and the level of drought stress for plants across a broad range of environments.
            For sixty-seven percent of the samples studied, Ψmin was measured as xylem or stem water potential. To obtain measures for Ψmin the leaves were covered with plastic and aluminum foil such that leaf and stem water potentials were equilibrated. In the remaining thirty-three percent of cases, Ψmin was measured as leaf water potential. In this case, Ψx may be less negative than leaf water potential. The water pressure drop across the leaf’s hydraulic pathway is caused by transpiration of water out of the leaves not covered by plastic and aluminum foil. The difference between Ψmin and Ψ50 represents the safety margin within which a plant is still able to function in its environment, and thus quantifies the degree of conservatism and ability to adapt to drought stress for a plant species’ hydraulic strategy. Plants that have a low safety margin are more prone to embolism and thus hydraulic failure. Because of the difference in leaf and xylem water potential, it is possible the amount of embolism in the stem could be overestimated. This overestimation would yield an overly narrow safety margin.
This study concludes that seventy percent of species in all forest biomes operate at narrow safety margins (Ψ88, still yielded similar results in vulnerability across biomes.

            Comparatively, gymnosperms had greater safety margins, and were therefore less susceptible to hydraulic failure. Forty-two percent of angiosperms operated at negative safety margins, in comparison to only six percent of gymnosperms. This discrepancy suggests that angiosperms have a greater resilience to reverse embolism by dissolving gases within the xylem, thus restoring the plant’s vascular system to a status where water can flow to photosynthetic sites. Still, this recovery can only occur if sufficient precipitation follows periods of drought in an ecosystem. This recovery process is therefore not effective in preventing die off if drought is severe and persists for a long period of time. Gymnosperms are not immune to hydraulic failure.
            The authors also found a strong association between
Ψ50 and MAP in the study. The mean and upper tenth quartile trends demonstrate significant decreasing resistance to embolism with increasing rainfall. The variation in evapotransporation (PET) and seasonality of precipitation: aridity index (MAP divided by PET) and mean precipitation of the driest quarter yielded similar results. This data was gathered from PET Global Aridity Index (Global-Aridity) and the Global Potential Evapo-Transpiration (Global-PET) Geospatial Database, respectively. For any climate region, there are various hydraulic strategies. Most variation in the Ψ50 is seen in sites with a MAP of 300 to 1,000 mm. High MAP sites (tropical rain forests) have less negative Ψ50, which suggests that low embolism resistance is related to low structural costs and high transport efficiency within the vascular system of plants. However, Ψ50 and MAP are not always correlated. In some cases, species can grow in more arid climates while still escaping water stress, therefore decreasing the need for high embolism resistance. Examples of such adaptations include riparian, groundwater-dependent vegetation, and drought-deciduous trees in tropical dry forests. Very negative Ψx are avoided by predictable access to groundwater through adaptations such as deep roots, internal water storage, and reduced leaf area. 
            Choat et al. suggest that most species operate close to their functional limits in regards to
Ψ50 and Ψmin. Woody species are susceptible to xylem failure during drought events, and are largely independent of rainfall region and biome. Plant species are able to adapt to drought to a certain extent, regardless of their ecosystem. However, there is a threshold at which subsequent mortality and die off will begin to occur due to xylem failure, embolism, and cessation of vital plant processes.
            Hydraulic strategies that operate within a “risky,” low safety margin to
Ψ50 gain a trade-off that balances growth with protection against risk of mortality within the environment. Stomatal behavior can take advantage of the range of xylem pressures within the hydraulic tolerance of a species, resulting in increased carbon gain. By regulating xylem pressure through the stomata to take advantage of various water availability scenarios, carbon uptake can be maximized. However, this behavior still places the plant at a greater risk of lost photosynthetic area or death.
            The link between a plant’s evolved embolism resistance and water availability is limited by long generation cycles of perennial plants. This adaptation capability raises important concerns regarding the capacity for plant species to adapt to the rapid pace of predicted climate change. If plant species are unable to keep pace with changing climates, this study suggests that net primary productivity will decrease, biodiversity will be lost, and the composition of ecosystems within forest will be altered. The authors note that although other factors induce drought-mortality (insect attack and carbohydrate depletion), these factors are still highly dependent upon one another. Shoat et al. conclude that embolism formation is the underlying mechanism of plant species die off and forest decline. Embolism sets thresholds for stomatal closure, limits photosynthesis, increases heat and light damage, and thus runs down carbohydrate reserves until a plant eventually dies. This study demonstrates that long-term monitoring of
Ψmin to quantify embolism resistance and hydraulic safety margins is necessary in accurately predicting the responses of forest ecosystems to climate change.

The Effects of Increasingly Severe Droughts on Carbon Sink Storage in North America

Temperate North America is currently a carbon sink, acting as a reservoir that accumulates and stores carbon. In less than one hundred years, this carbon sink could disappear due to projected changes in drought severity and frequency caused by anthropogenic global climate change. Based on comparisons of the Palmer Drought Severity Index and paleoclimate reconstructions of droughts from tree ring data, the drought that lasted between the years 2000 and 2004 in western North America was the most severe to occur in eight hundred years. According to Schwlam et al. (2012), the carbon dioxide absorption ability of the western North American carbon sink declined from 30 to −298TgCyr−1 in 2000–2004, due to the effects of drought on the ecosystems’ carbon cycling functions. This study suggests that the western North American carbon sink that currently stores 177–623TgCyr−1 could be lost in less than one hundred years, due to the increasing severity of droughts. —Hilary Haskell
Schwalm, CR., 2012. Reduction in Carbon Uptake during Turn of the Century Drought in Western North America. Nature 5: 55156.

            Schwalm et al. quantified drought with precipitation, soil moisture, evaporative fraction and latent heat data. Data from the U.S. Geological Survey for the five main river basins in the western United States during the years 1997–2007 indicated that water availability decreased. The National Agricultural Statistics Survey for 2,383 counties in the western United Sates suggested that cropland productivity declined 5% during the 2004–2005 drought. According to precipitation predictions, the drought in North America between the years 2000–2004 will become a typical occurrence. These hydroclimatic predictions were reached using reliable methods. Summer precipitation and summer PDSI are two separate indices; however, linking the PDSI-based analysis of past drought events with summer precipitation predictions for the future is a well-supported practice. The validity of these assumptions is supported by the Coupled Model Intercomparison Project Phase 5. Precipitation levels and tree-ring PDSI exhibited highly similar frequency distributions between the data sets compared, indicating consistency and accuracy for the PDSI. The similarities between all three drought indicators demonstrate that they are able to reliably predict the frequency and severity of drought. Not only do drought indices indicate increased aridity of western North America, but also trends in snowpack decline. Although regional precipitation is difficult to forecast, climate model predictions generally underestimate drought extent and severity.
            The authors observed carbon and energy fluxes at fifteen eddy-covariance flux tower sites that are part of the global FLUXNET network, using data from the North American Carbon Project Site Synthesis and Ameriflux. This study only considered an eddy-covariance flux tower site in western North America if it provided one year of data for both drought and non-drought conditions in 1997–2007. Across these fifteen sites, despite differences in ecosystems, time period, soils, climates, and other ecosystem disturbances, there was still an obvious decrease in carbon uptake. A decrease in net ecosystem productivity (NEP) of −63gCm2yr1with a 95% confidence interval:−20 to −139gCm2yr1 indicated the resulting decrease in carbon sink storage.
            For grasslands and evergreen needleleaf forests, the drought resulted in reduced gross primary productivity (GPP) that outweighed the effects of ecosystem respiration, creating an overall reduction in carbon dioxide uptake. Respiration is the process by which plants convert the sugars produced during photosynthesis back into energy so that they may be metabolized, resulting in water and carbon dioxide byproducts. Ecosystem respiration corresponds to the sum of all plant respiration in an ecosystem. Conversely, woody savannas demonstrated an increase in carbon dioxide uptake due to reduced ecosystem respiration from the reduced rate of decomposition during droughts.
            The authors also analyzed latent heat flux (LE) as a determinant of drought for three land cover classes (grassland, evergreen needleleaf forests, and woody savannas), and found differing magnitudes of decreased LE. LE is the measure of the exchange of heat between the Earth’s surface and the atmosphere, due to evaporation and transpiration into the troposphere and condensation of water out of the troposphere. LE is used as an indication of drought.  Evergreen forests demonstrated the greatest decrease in LE, followed by woody savannas and needleleaf forests. Sensible heat flux (H), the conductive heat flux between the atmosphere and Earth’s surface (measured by eddy covariance), was less variable than LE. But, there was an increase in H for needleleaf forests and woody savannas that also had a temperature increase of 0.40°C and 0.41°C,respectively, from JuneSeptember. However, H decreased in grasslands due to increased albedo effects from dieback that exposed darker soil and turned bright leaf tissue darker.
            Drought-induced regional deviations for water and carbon balances were seen in precipitation, soil moisture, instrumental era PDSI, Moderate Resolution Imaging Spectoaiometer (MODIS) GPP, MODIS net primary productivity (NPP) and empirically extrapolated FLUXNET data obtained during the years 19972007. Regionally, Montana and Idaho demonstrated the largest changes relative to the baseline period years: 19971999 and 20052007. During these years, area-averaged soil moisture declined 4mm per month, with the largest reduction of 45% occurring during the summer. As expected, area-averaged precipitation decreased 6 mm per month. On the PDSI scale, negative values indicate periods of drought. Decreased precipitation can be seen in the 0.5 baseline period value for the PDSI scale falling to 1.6 for the average value in the region. The largest change of 1.6 to 3.6 on the PDSI scale indicated a shift from slightly wet conditions to a severe drought in the region. During the 20002004 drought, 75% of the sites studied demonstrated mild to severe drought.
            By combining the effects of drought over the region, the authors concluded that there was a considerable reduction in region-wide productivity and carbon uptake. Schwalm et al. quantified the decrease in GPP by −182TgCyr−1 (−38gCm−2yr−1) for  extrapolated FLUXNET of monthly NEP, GPP, and ecosystem respiration; and −234TgCyr−1(−47gCm−2yr−1) for MODIS measures of NPP and GPP.  Both the FLUXNET data and MODIS estimates of NPP and GPP were forced with reanalysis, remote sensing, meteorological, and land cover data. When evaluating the change in the carbon sink strength, only the differences due to the drought that lasted from 20002004 in comparison to the baseline period from 1997−1999 and 20052007 were considered. The three FLUXNET-based estimates of drought anomalies along with the two estimates based on inversions (Jena CO2  Inversion and Carbon Tracker were used as independent estimates of the land sink during the years 19972007)  indicate that the 20042005 western North America drought caused a decline in terrestrial carbon sink strength (NEP)  ranging from −30 to −298TgCyr−1, relative to a baseline sink strength of 177–623TgCyr−1. This reduced the uptake of carbon dioxide by an average of 51%.
            In order to substantiate the severity of the drought in western North America during the years 20042005, the authors considered the paleoclimatic tree ring record to reconstruct the PDSI index for the years 8002006. Doing so revealed that the drought considered in this study had the lowest PDSI value for a five year drought in eight hundred years. In the 1,207 years of paleoclimate data surveyed, there were only two drought events of comparable severity, and both occurred during periods of historical megadroughts that lasted much longer than the drought in this study. On average, these megadroughts reflected less drought severity and were more geographically limited in scale. When only considering single-year summer PDSI values, over the course of the last 2,000 years, tree ring data indicates that only ninety-seven summers were as or more severe than the drought that lasted between 20002004.
            Based on forecasted precipitation patterns and the recent continuation of decreased levels of precipitation, western North America is headed towards a twenty-first century megadrought. This megadrought will further decrease crop productivity, primary production in ecosystems, LE, runoff to water basins, and carbon dioxide uptake, and could possibly cause these patterns to become permanent. The effects of megadrought in the future could greatly diminish the North American carbon sink’s carbon accumulation and storage potential if global climate change patterns continue as projected. 

Predicting Ecosystem Resilience and Drought-Tolerance Thresholds in the face of Global Climate Change and Drought

Global temperatures during the years of 2000–2009 were some of the driest and warmest on record. These changing climatic conditions reflect the importance of ecosystem resilience in the face of altered temperature and precipitation patterns. Ecosystem resilience is the ability of ecosystems to function with the same feedback loops and sensitivity to changed surroundings during periods of disturbance as they would under undisturbed conditions. This function has implications for vegetation productivity, and therefore, carbon balance and food security as well. Ponce Camps et al., (2012) measured ecosystem resilience as the capacity of ecosystems to absorb disturbances from early twenty first century drought, while still maintaining late twentieth century above ground net primary production (ANPP) when there was high annual water availability. The authors found that ecosystem water use efficiency (WUEe) was similar across biomes, suggesting that ecosystems are able to cope with variations in hydroclimatic conditions by adjusting to both droughts and periods with high levels of precipitation, but only up to a point. Future climate change and predicted increases in drought frequency and severity may well overwhelm his ability. The authors compared data from both hemispheres over the years 2000–2009 to data from 1975–1988, in order to predict future ecosystem resilience and the threshold for WUEe. —Hilary Haskell
            Ponce Camps, G., Moran, S., Huete, A., Zhang, Y.,  Bresloff, C., Huxman, T., Eamus, D., Bosch, D.D.,  Buda, A.R., Hearstill Scalley, T., Kitchen, S.G., McClaran, M.P., McNab, W.H., Montoya, D.S., Morgan, J.A., Peters, D.P.C., Sadler, E.J., Seyfried, M.S., Starks, P.J., 2012. Ecosystem Resilience despite Large–scale Altered Hydroclimatic Conditions. Nature.

Ponce Campos et al. used a data set for this study that consisted of twelve U.S. Department of Agriculture sites and seventeen sites in Australia, including a variety of rainfall levels from both the Northern and Southern hemispheres. This data set from the years 2000–2009 was compared to data from 1975–2008 as a baseline to demonstrate the change in hydroclimatic conditions over recent years. The 1998 data came from long term ecological research sites. Precipitation and temperature were measured in a homogeneous vegetated area with no major disturbances in 2000–2009.
To quantify ecosystem resilience, the authors used the ecosystem’s functional response to disturbances characterized by rain–use efficiency (RUE) and water–use efficiency (WUEe). RUE is calculated by dividing ANPP by precipitation over a period of time. Similarly, ecosystem water–use efficiency was calculated by dividing ANPP by evapotranspiration. Evapotranspiration is defined as precipitation minus the water lost from plant leaves. The drought events of the early 2000s are recognized as a departure from typical climate variability. There was significant (P < 0.002) decrease in the Palmer Drought Severity Index (PDSI) over 1980–1999 and 2000–2009 for the U.S. and Australia, indicating increased aridity. Also, warm season temperatures over the 2001–2009 time period were significantly higher, (P
The Moderate Resolution Imaging Spectroradiometer (MODIS) produced the Enhanced Vegetation Index (EVI) through satellite observations to estimate collective plant behavior. This study quantified the relationship for the biomes and precipitation patterns between the EVI data and ANPP field methods for 2009–2009, using data from ten sites in the United States, and developed the following relationship: ANPPs = (51.42) EVI1.15. From this equation, there is evidence that the RUE data from the late twentieth century are consistent with plant production responses to precipitation. Low mean RUE for biomes with high precipitation suggest that some water from precipitation is not consumed by plants, but rather, abiotic aspects of the water cycle. This finding is further confirmed by comparing the positive relationship between evapotranspiration and plant production. Ponce Campos et al. concluded that the mean ecosystem water use efficiency was constant across the entire precipitation gradient. Furthermore, there were no significant differences among WUEe for the three data sets, suggesting that biomes remain sensitive to water availability during warmer drought conditions.
The most severe drought years coincided with maximum ecosystem WUEe in all biomes studied, indicating that ecosystems remain productive in extremely dry years by increasing their WUEe. However, Ponce Campos et al. also considered ecosystem resilience during the wetter mid-to-late drought (2003–2009) years, compared to hydroclimatic conditions in 1975–1998, and found that WUEe across all biomes and hydroclimate periods was at a minimum value. WUEe did not significantly (P > 0.05) vary in different hydroclimate periods, suggesting that biomes are able to respond to high annual precipitation, even during drought, by using available water more efficiently.
WUEe can increase or remain constant during changing hydroclimate patterns. During drought, the authors found that biomes’ WUEe increased as drought severity increased. Therefore, plant productivity remained near constant at late–twentieth century levels despite decreased precipitation. However, during wetter years, the biomes demonstrated similar consistency through absorbing and adapting to the disturbance of drought, and remained sensitive to water availability based on ANPP.
Ponce Campos et al. recognize that there are additional factors that may account for the plants’ and biomes’ responses to precipitation disturbances, such as vegetation structure and function and plant-soil feedbacks that the RUE or WUEe do not take into account. These factors are also important in considering ecosystem’s vulnerability and tolerance to changing hydroclimatic patterns.
During dry years, the sites with high plant productivity also had higher WUEe. These data indicate that vegetation is able to remain sensitive to water availability, coping with the stress of drought or respond to high annual precipitation with extra growth. This adaptability therefore demonstrates an ecosystem’s resilience to more severe and frequent changes in hydroclimatic conditions, from global climate change. During drought, ecosystems with high plant productivity during normal hydroclimatic had WUEes that were similar to the plants adapted to less productive, arid ecosystems. However, the authors warn that if all ecosystems are subjected to limited water variability, there will be a cross-biome maximum WUEe that will not be sustainable if drought continues and hydroclimatic conditions continue to become drier and hotter.
The ANPP/evapotranspiration model from cross–biome WUEe will not be sustainable if more arid, hot hydroclimatic conditions persist. These changing hydroclimate conditions will surpass a threshold that biomes are able to endure, without resulting in drought–induced mortality. Die off and decreased resilience would occur for ecosystems with the most variability in precipitation (grasslands). For grasslands, increasing aridity during prolonged warm drought led to a decrease in WUEe and resilience.
Ecosystems are able to respond and adapt to droughts, by an increase in WUEe during dry years and through resilience during wet years, in order to maintain vegetation productivity. This study combines data gathered from several biomes over different time periods on two continents, with changing hydroclimatic conditions. Ponce Campos et al. were able to find that in the face of global temperature increases and the increased frequency and severity of droughts, there will be increased vegetation mortality that could threaten ecosystems’ resilience across biomes. This decrease in resilience would result from species die off and thus changes in ecosystem structures. Thus, species die off could negatively affect crop yields and soil carbon balance feedback loops in the future. 

The Effects of Drought and Climate Change on the Resilience and Function of Grasslands based on Grass Species’ Drought Tolerance

Grassland ecosystems comprise a large proportion of the terrestrial biosphere, and are composed of 11,000 different types of grass species. Due to the increasing frequency and severity of global droughts and climate change, there will be implications for plant productivity, geographic species distribution, and widespread plant mortality, especially for the grass species of grassland ecosystems. Disturbances such as distribution, productivity, and productivity can change species composition of grassland ecosystems. Mortality of some species could affect the overall function and resilience of these grassland ecosystems. This study investigates physiological drought tolerance of grass species and its implications across phylogenies and climates in the context of the resilience and function of grassland ecosystems, during and after periods of drought (Craine et al., 2013). An ecosystem’s functions can remain unchanged by drought, if native drought tolerant grass species are already there, and can increase in relative abundance. If not, non-local drought-tolerant grass species may invade, and change the ecosystem functionality, for example by having higher rates of water and carbon dioxide exchange than before.—Hilary Haskell
            Joseph, Craine M., Troy W. Ocheltree, Jesse B. Nippert, Gene Towne, Adam M. Skibbe, Steven W. Kembel, and Joseph E. Fargione. “Global Diversity of Drought Tolerance and Grassland Climate-change Resilience.” Nature Climate Change 3 (2013): 63-67.


            Craine et al. researched the physiological drought tolerance and leaf functional traits for 426 grass species. The grass species used in this study originated from six different continents, and were grown from seeds either provided by the United States Department of Agriculture or hand-collected in New Zealand. The grasses were grown in a Conviron Growth Chamber, where temperatures were maintained at 25 °C for sixteen hours and 20°C for eight to simulate night and day conditions. Plants received water on a daily basis and fertilizer biweekly. The main parameter for this study was critical leaf water potential, which the authors defined as the level at which stomata conductance of gases such as water vapor, carbon dioxide, and oxygen flow through the stomata of leaves falls below an ecological threshold for continued function and plant vitality. To find critical leaf water potential, five weeks after germination or planting of the grass species studied, the authors stopped watering one sample of each species to simulate a drought. Then, the authors used a steady-state diffusion porometer to measure leaf conductance daily until stomatal closure occurred. After closure, Craine et al. measured the hydrostatic pressure potential, the energy per unit volume of water exerted by the pressure of overlying water, using a Scholander pressure bomb. The leaf water potential represents the tension on the water column of the plant, with lower values indicating greater physiological tolerance of plants to dry soils. From a specie’s dry soil tolerance, the authors inferred that a species was also tolerant of drought. The leaf water potential of stomatal closure is the critical leaf water potential, Ψcrit, for the species.

The diversity of drought tolerant species across global climates is unknown. However, grasslands with a diversity of local drought tolerant grass species are able to adapt to drought without altering overall ecosystem function more so than grasslands with less local drought tolerant grass species diversity. The functions of these grassland ecosystems include carbon uptake, productivity, soil retention, and provision of forage to grazers. These functions could change if there is a lack of diversity across the types of drought tolerant grasses that support these functions, and a subsequent die off of these local drought intolerant species. However, for non-diverse drought tolerant ecosystems, if non-local drought tolerant grass species migrate to the ecosystem during a drought, these functions could be maintained, albeit ecosystem function may be altered post-drought due to the presence of the non-local grass species. Non-local drought tolerant grass species may have different functional traits than the local drought intolerant grass species initially present, thus changing the functional composition of the ecosystem.
 Based on some predictions for the rate of climate change in coming years, grass species, and thus ecosystem functions, will not be able to adapt quickly enough to projected variations in temperature and precipitation. After periods of drought in grasslands with low drought tolerant grass species diversity, if non-local drought tolerant grass species migrate to these grassland ecosystems, this could alter the functional composition of ecosystems and the processes that these ecosystems carry out. Craine  et al., note that currently, there is not enough data to predict the impacts of drought on the functional composition of grasslands.  
            Grass species’ drought tolerance varied across the range of mean annual precipitation (MAP) for grasslands (250-1,500mm). This variance is important to the ecosystem’s functional composition and resilience, due to increasingly frequent and severe drought patterns. The highest and lowest Ψcrit for species within 50 mm precipitation intervals of MAP both increased significantly with increasing MAP. The median Ψcrit was (leaf water potential), −4.1MPa and ranged between −1.4 MPa to −14MPa. Climate envelopes, areas with prevailing meteorological conditions such as precipitation and temperature, were created for 52% of the species studied. For the species in climate envelopes, Craine et al. were able to test the grass’s MPa and the upper and lower bounds of Ψcrit. Ψcritmaximum and minimums only shifted by 1.0MPa, even though wetter regions had more drought intolerant species than drought tolerant species. There were no shifts in the minimum and maximum Ψcrit for the 10% and 90% ranges of MPa, and the inner quartile were only 0.5 MPa or less from MAP 250−1,500mm of precipitation.
            Based on these findings, only humid ecosystems with very high levels of precipitation (MAP>1,500mm) would experience changes in functional ecosystem composition and thus functional responses, due to the lack of diversity of drought tolerant species initially local to the ecosystems. There is a negative correlation between maximum drought tolerance and increasing MAP for ecosystems with MAP greater than 1,913 mm. Furthermore, for the ecosystems with higher levels of precipitation, maximum drought tolerance declined with increasing MAP at a faster rate than ecosystems that received less rain. Although these findings are significant, Craine et al., acknowledges that there were relatively few species tested (only about 10% of the species sampled could be used to create climate envelopes). Furthermore, temperature gradients did not have a significant effect on drought tolerance of grassland species. This study suggests that ecosystems with high levels of precipitation might require species migration of non-local drought tolerant species to maintain ecosystem function, due to the lack of diverse drought tolerant species originally local to the ecosystem.
            Craine et al. also examined the variation in physiological drought tolerance for grass species in a single grassland, the Konza Native Tallgrass Prairie in northeastern Kansas. The data from this grassland was used for comparison to global data on physiological drought tolerance of grassland species. On the Prairie, mean average temperature is 13 °C and average monthly and temperatures ranged from -3 °C to 27°C. Annual rainfall averaged 833 mm over the years 19832009.  Fifty-two of the 426 species used in this experiment were collected from the Konza Prairie. The authors concluded that local variation of physiological drought tolerance of grass species in grassland ecosystems is high. Craine et al. compared the
Ψcrit distribution from the Konza species with global distribution to find that the global range of physiological drought tolerance for grass species was present at this site alone.
            In addition to comparing geographic drought tolerance of grass species, Craine et al. also compared phylogentic variations in drought tolerance using taxonomic data for species from the uniprot database. The authors hypothesized that if certain clades of the same ancestor were to die off, the ecosystem should still maintain its functional diversity of physiological drought tolerance. By studying sixty-five grass species classified into various phylogenies, the authors found that physiological drought tolerance had evolved numerous times, and was widespread in the grass phylogeny. According to the Ψcritbetween clades, there was no significant difference (BEP versus PACMAD; P=.016) or the four subfamilies sampled. Furthermore, C4 species were not on average more or less physiologically drought tolerant than C3species.
            Drought tolerant grass species (
Ψcrit <−4.1 MPa) have unique traits in comparison to drought intolerant grass species (Ψcrit>−4.1MPa). Craine et al. used recently expanded leaves and a Li 6400 infrared gas analyzer with red/blue emitting diode light source and CO2 injector to find that drought tolerant species had higher photosynthetic rates (17.0±0.5 versus 15.3±0.5 μmolm−2s−1, P=0.01), and higher stomatal conductance (0.18 5± 0.008 versus 0.152 ± 0.008molm−2s−1, P=0.002) compared with physiologically drought-intolerant grass species (Ψcrit>−4.1MPa). This finding suggests that evolution of grass species caused by droughts could result in the increased abundance of species that are able to re-adapt to higher levels of precipitation through increased primary productivity post-drought. Higher gas-exchange rates per-unit leaf area is one of the main functional traits that contribute to this finding. There seemed to be a tradeoff between leaf widths and drought tolerance for the 426 grass species examined. The tradeoffs created a triangular model for the functional and structural traits of grasses: wide leaved and drought intolerant, narrow-leaved drought intolerant, and narrow-leaved drought-tolerant. Interestingly, there were no drought-tolerant species with wide leaves. According to Craine et al., drought adaptations tend to be structural rather than morphological. In addition, of twenty species studied, drought-tolerant grasses had narrower xylem, the supporting tissue of vascular plants that conducts water through tracheids and vessels.
            Craine et al. conclude that physiological drought tolerance in the grass species of grassland ecosystems is a widely, commonly evolved trait that is distributed both phylogenetically and geographically. Because of this broad distribution of drought tolerance in a diversity of grassland bioclimates, species local to the ecosystem can support natural ecosystem function when drought occurs, without the migration of non-local grass species to the ecosystem. In the face of decreased precipitation, this study suggests that grasslands with high functional diversity for drought tolerance will be resilient to drought without immigration of non-local drought tolerant grass species to maintain ecosystem function. To study ecological survival strategies for drought intolerant species, the authors used the triangle relationship of physiological drought tolerance versus leaf width. This triangular relationship of structural and morphological function includes drought tolerance and leaf width; however, it still leaves out other ecological factors such as shade tolerance and nutrient availability. Furthermore, grass species in areas with low precipitation had significantly (P
            The effects of climate change and drought on grassland communities still requires more research. However, according to Craine et al., grassland ecosystems across bioclimatic gradients with grass species that are functionally diverse in their drought tolerance should be able to adapt to drought in order to retain ecosystem function in the face of climate change. The diversity of the drought tolerant grass species local to a grassland ecosystem is imperative to the ecosystem in adapting to more frequent, severe droughts that are predicted to increase with climate change in the future. With the findings from this study, the authors are better able to promote resilient ecosystems and model the effects of droughts on grassland ecosystems.