Impacts of Land–Atmosphere Feedbacks on Deep, Moist Convection on the Canadian Prairies

Julian C. Brimelow Centre for Earth Observation Science, Department of Environment and Geography, University of Manitoba, Winnipeg, Manitoba, Canada

Search for other papers by Julian C. Brimelow in
Current site
Google Scholar
PubMed
Close
,
John M. Hanesiak Centre for Earth Observation Science, Department of Environment and Geography, University of Manitoba, Winnipeg, Manitoba, Canada

Search for other papers by John M. Hanesiak in
Current site
Google Scholar
PubMed
Close
, and
William R. Burrows Cloud Physics and Severe Weather Research Section, and Hydrometeorology and Arctic Lab, Meteorological Research Division, Science and Technology Branch, Environment Canada, Edmonton, Alberta, Canada

Search for other papers by William R. Burrows in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

The purpose of this study was to focus on how anomalies in the normalized difference vegetation index (NDVI; a proxy for soil moisture) over the Canadian Prairies can condition the convective boundary layer (CBL) so as to inhibit or facilitate thunderstorm activity while also considering the role of synoptic-scale forcing. This study focused on a census agricultural region (CAR) over central Alberta for which we had observed lightning data (proxy for thunderstorms), remotely sensed NDVI data, and in situ rawinsonde data (to quantify impacts of vegetation vigor on the CBL characteristics) for 11 summers from 1999 to 2009. The authors’ data suggest that the occurrence of lightning over the study area is more likely (and is of longer duration) when storms develop in an environment in which the surface and upper-air synoptic-scale forcing are synchronized. On days when surface forcing and midtropospheric ascent are present, storms are more likely to be triggered when NDVI is much above average, compared to when NDVI is much below average. Additionally, the authors found the response of thunderstorm duration to NDVI anomalies to be asymmetric. That is, the response of lightning duration to anomalies in NDVI is marked when NDVI is below average but is not necessarily discernible when NDVI is above average. The authors propose a conceptual model, based largely on observations, that integrates all of the above findings to describe how a reduction in vegetation vigor—in response to soil moisture deficits—modulates the partitioning of available energy into sensible and latent heat fluxes at the surface, thereby modulating lifting condensation level heights, which in turn affect lightning activity.

Corresponding author address: Dr. John M. Hanesiak, Centre for Earth Observation Science (CEOS), 468 Wallace Building, University of Manitoba, Winnipeg MT R3T 2N2, Canada. E-mail address: john_hanesiak@umanitoba.ca

Abstract

The purpose of this study was to focus on how anomalies in the normalized difference vegetation index (NDVI; a proxy for soil moisture) over the Canadian Prairies can condition the convective boundary layer (CBL) so as to inhibit or facilitate thunderstorm activity while also considering the role of synoptic-scale forcing. This study focused on a census agricultural region (CAR) over central Alberta for which we had observed lightning data (proxy for thunderstorms), remotely sensed NDVI data, and in situ rawinsonde data (to quantify impacts of vegetation vigor on the CBL characteristics) for 11 summers from 1999 to 2009. The authors’ data suggest that the occurrence of lightning over the study area is more likely (and is of longer duration) when storms develop in an environment in which the surface and upper-air synoptic-scale forcing are synchronized. On days when surface forcing and midtropospheric ascent are present, storms are more likely to be triggered when NDVI is much above average, compared to when NDVI is much below average. Additionally, the authors found the response of thunderstorm duration to NDVI anomalies to be asymmetric. That is, the response of lightning duration to anomalies in NDVI is marked when NDVI is below average but is not necessarily discernible when NDVI is above average. The authors propose a conceptual model, based largely on observations, that integrates all of the above findings to describe how a reduction in vegetation vigor—in response to soil moisture deficits—modulates the partitioning of available energy into sensible and latent heat fluxes at the surface, thereby modulating lifting condensation level heights, which in turn affect lightning activity.

Corresponding author address: Dr. John M. Hanesiak, Centre for Earth Observation Science (CEOS), 468 Wallace Building, University of Manitoba, Winnipeg MT R3T 2N2, Canada. E-mail address: john_hanesiak@umanitoba.ca

1. Introduction

Brimelow et al. (Brimelow et al. 2011) presented findings from a novel study designed to explore linkages between the normalized difference vegetation index (NDVI) and lightning duration (DUR) for 38 census agricultural regions (CARs) on the Canadian Prairies. Brimelow et al. (Brimelow et al. 2011) found evidence for coupling between NDVI and DUR during the summer months over drought-affected regions on the Canadian Prairies, with drought conditions associated with less lightning activity. They also noted that the strength of the relationship between NDVI and DUR increased significantly as both the area and magnitude of the dry anomaly increased. In contrast, they noted a very weak relationship between NDVI and DUR over those CARs experiencing pluvial conditions. They concluded that these observations suggest that dense vegetation alone is not a necessary or sufficient condition for above-average lightning duration. Here we focus on a CAR over central Alberta to further elucidate how anomalies in NDVI can condition the convective boundary layer (CBL) so as to inhibit or facilitate thunderstorm activity while also considering the role of synoptic-scale forcing on modulating summer thunderstorm activity.

Modeling studies have suggested that changes in low-level moisture and energy fluxes associated with soil moisture and vegetation anomalies can perpetuate drought conditions (e.g., Dirmeyer 1994; Beljaars et al. 1996; Seneviratne et al. 2010). Vegetation is known as an important pathway for coupling the surface and the atmosphere (e.g., Chen and Zhang 2009). As a result, changes in vegetation vigor can have a marked impact on the partitioning of incoming solar radiation (e.g., Liu et al. 2006) and on modulating surface fluxes (e.g., Dominguez and Kumar 2008), which can in turn affect thunderstorm activity (e.g., Toumi and Qie 2004; Taylor et al. 2010). Consequently, coupling between the land surface and overlying atmosphere is strongest during the warm season (Koster et al. 2004). The strength of the land–atmosphere interaction, or coupling, is difficult to quantify. For example, Dirmeyer (Dirmeyer 2006) suggested that soil moisture is “strongly” modulated by antecedent precipitation and that soil moisture in turn exerts a “moderate” influence on evapotranspiration (ET) but that the impact of ET on precipitation is “tenuous.” Modeling studies (Koster et al. 2004; Koster et al. 2006) have identified “hot spots” where coupling between the land surface and atmosphere is marked. The Great Plains of North America, including southern portions of the Canadian Prairies, is one such hot spot. Dirmeyer et al. (Dirmeyer et al. 2009) used observations and atmospheric reanalyses data to conduct an integrated analysis for quantifying land–atmosphere interactions, and this too identified the Great Plains and southern Canadian Prairies as hot spots for land–atmosphere coupling in the summer.

Using observed surface temperature and moisture, soil moisture, and precipitation data, Findell and Eltahir (Findell and Eltahir 1999) did not find a positive correlation between the surface moist static energy (MSE) and rainfall over Illinois. There was, however, a significant negative correlation between the soil moisture (top 10 cm) and the lifting condensation level (LCL) and between the LCL and the rainfall. Alfieri et al. (Alfieri et al. 2008) investigated the relationship between warm season soil moisture and subsequent precipitation in the U.S. Midwest from 1971 to 2003. They identified convective precipitation days using convective available potential energy (CAPE) calculated from observed soundings. Next, they correlated convective rainfall incidence against soil moisture (top 50 cm) with precipitation lagging soil moisture by one day. They concluded that the possibility of a positive feedback between soil moisture and convective precipitation exists but that soil moisture alone cannot unambiguously isolate the signal.

Recently, satellite-derived vegetation indices and soil moisture data have been used to try to determine the effects of changes in vegetation vigor during the warm season on the boundary layer and subsequent precipitation. However, the results have not been conclusive. Notaro et al. (Notaro et al. 2006) used photosynthetically active radiation (PAR) data derived from satellite data to examine the interaction between vegetation and precipitation over the United States between 1982 and 2000. They noted that the impacts of PAR on precipitation were complex and weaker than the feedback for temperature. For example, the correlation for PAR leading precipitation was negative over the northern Great Plains in June–August (JJA) but positive in fall and spring. Liu et al. (Liu et al. 2006) conducted a similar study to that of Notaro et al. (Notaro et al. 2006) but for all ice-free landmasses in the world. Specifically, they investigated feedbacks between PAR, observed surface temperature, and precipitation. Liu et al. found that, over the boreal regions of Eurasia and North America, there is a positive feedback between PAR and temperature on account of the surface albedo feedback. Over the Canadian Prairies, they identified a negative correlation when PAR led precipitation by one month during the summer months, which suggests a negative feedback between precipitation and PAR there.

Wang et al. (Wang et al. 2006) applied Granger causality tests to observed gridded temperature, precipitation, and NDVI from 1982 to 2000 over the northern grasslands of the United States. They found a significant causal relationship between NDVI early in the season and precipitation and temperature later in the season (July onward). They purport that initially enhanced vegetation may deplete soil moisture faster, thereby leading to drier and warmer anomalies later in the season. Kim and Wang (Kim and Wang 2007) investigated the soil moisture–vegetation–precipitation feedback over the Mississippi River basin using ensemble simulations from the coupled Community Atmosphere Model–Community Land Model. They focused on the impact of vegetation feedbacks arising from initial soil moisture anomalies through the entire soil column and found that, in JJA, initial wet soil moisture anomalies increased precipitation through increased ET.

One possibility for these conflicting results is that many of them were based primarily on numerical modeling output. For example, Hohenegger et al. (Hohenegger et al. 2009) found that the sign of the soil moisture–precipitation feedback in model simulations over Europe depended on the horizontal grid spacing and cloud convection schemes used in their model. Kim and Wang (Kim and Wang 2007) found that initial dry soil moisture anomalies did not show a significant impact on precipitation and vegetation but concluded that this was because of the dry bias in the model they used. Additionally, studies examining feedbacks between soil moisture and vegetation anomalies and precipitation often do not discern between convective and stratiform precipitation.

Theoretical and modeling studies have shown that the amount of moisture in the CBL has important consequences for the initiation and intensity of thunderstorms (Segal et al. 1995; Crook 1996; Yamada 2008). Also, conceptual models describing the physical mechanisms whereby anomalies in the underlying soil moisture and vegetation can affect subsequent thunderstorm activity mostly have their origins in modeling studies (e.g., Trier et al. 2004; Findell and Eltahir 2003a; Findell and Eltahir 2003b; Pal and Eltahir 2001; Schär et al. 1999). In contrast, here we use mostly observations to examine the relationship between the land surface and thunderstorms.

Given that convection is an important component of the hydrological cycle on the Canadian Prairies (Raddatz and Hanesiak 2008), the purpose of this study was to focus on how anomalies in NDVI can condition the CBL so as to inhibit or facilitate thunderstorm activity while also considering the role of synoptic-scale forcing in modulating summer thunderstorm activity. Our study focused on a census agricultural region over central Alberta (CAR11) for which we had observed lightning data (proxy for thunderstorms), NDVI data (measure of vegetation vigor and a proxy for soil moisture), and in situ rawinsonde data for 11 summers (1 June through 31 August) from 1999 to 2009. The rawinsonde data were used to assess the response of the CBL to changes in NDVI. The study region (Figure 1) is also of interest because it is likely located within a land–atmosphere feedback hot spot (e.g., Koster et al. 2006).

Figure 1.
Figure 1.

Map of the study area. The polygon over central Alberta is CAR11, which is inside the agricultural area (indicated by the green border). The red dot represents the location of the Stony Plain upper-air site.

Citation: Earth Interactions 15, 31; 10.1175/2011EI407.1

The paper is structured as follows: In section 2, the various datasets and our methodology are described. In section 3, the relative roles of surface and upper-air forcing and NDVI on lightning duration are explored. Section 4 focuses on the relationships between observed sounding parameters and NDVI and how these are related to lightning duration. In section 5, we present a conceptual model that integrates our findings concerning the linkages between NDVI and CBL characteristics, synoptic-scale forcing, and lightning duration. The paper concludes with a summary of our primary findings.

2. Data and methodology

2.1. Study area

Statistics Canada divides the agricultural zone on the Canadian Prairies into CARs for the purpose of calculating agricultural statistics. Here we will consider relationships between vegetation vigor, synoptic-scale forcing, and lightning duration over CAR11 located in central Alberta (see Figure 1). CAR11 has an area of about 16 500 km2 (equivalent radius of ~72 km) and is located in the parkland ecoclimatic zone, which marks the transition zone between the boreal forest and the prairie grasslands, and is thus representative of vegetation type over a large portion of the Canadian Prairies. Western portions of the CAR are located about 100 km northeast of the Rockies, with the centroid located about 200 km northeast of the Rockies. Terrain heights range from 650 m (MSL) in the northeast to 950 m in the southwest. About 60%–65% of CAR11 is dedicated to agriculture (crops and pasture), with pockets of Aspen groves, especially in the far west, north, and east. Crops are predominantly cereal crops. Planting dates vary from year to year and range from early May to the end of May, depending on soil moisture and temperature conditions, with crops typically maturing in mid-August (e.g., Mkhabela et al. 2010). NDVI data between 1987 and 2009 show NDVI values over CAR11 peaking in the week of 19–25 July with a rapid decline after mid-August.

2.2. Cloud-to-ground lightning data

We used cloud-to-ground (CG) lightning flash data from the Canadian Lightning Detection Network (CLDN) in this study as a proxy for thunderstorm occurrence and longevity over the study area. Detection efficiency of CG flashes for the CLDN over the prairie agricultural zone is >90%, whereas the median stroke location accuracy is 500 m or better (Cummins et al. 1999).

Thunderstorm activity is quantified using the lightning duration within CAR11. The lightning duration is defined as the length of time (in hours) between 0930 and 1930 local solar time (LST) where flashes occurred “continuously”; during the summer months in central Alberta, LST typically precedes UTC by about 7.5 h. No more than 10 min was allowed to separate a flash and the next flash in a given series of flashes. If more than 10 min separated two flashes, then the time series was discontinued and a new time series was started when the next flash was recorded. The lengths of the series for each solar day were summed to calculate monthly and seasonal totals.

We used the 0930–1930 LST (1700–0300 UTC) time window to focus on surface-based thunderstorms (rooted in the convective boundary layer) that typically develop in response to daytime heating rather than nocturnal thunderstorms that tend to be elevated. According to Burrows and Kochtubajda (Burrows and Kochtubajda 2010), from 1999 to 2008, CG lightning within a 50-km radius of Edmonton International Airport was a minimum between 1200 and 1600 UTC with the peak time for lightning activity occurring between 2200 and 0400 UTC. From 1999 to 2008, about 60% of all lightning strikes over CAR11 in JJA were observed between 1700 and 0300 UTC.

2.3. NDVI

We used NDVI as a measure of vegetation vigor and a proxy for root-zone soil moisture content. Vegetation is known as an important pathway for coupling the surface and the atmosphere, and vegetation health affects the partitioning of available energy and moisture flux into the CBL. The latent heat flux of moisture is an important source of moisture in the CBL, with transpiration from a dense and unstressed canopy under favorable conditions capable of increasing the mixing ratio of the CBL by 4–8 g kg−1 day−1 (Raddatz 1993; Segal et al. 1995).

NDVI (see Tucker 1979) is the normalized difference between the near-infrared and visible red reflectance, and it responds to changes in both the chlorophyll content and the intracellular spaces in the spongy mesophyll of plant leaves (Gu et al. 2007). Higher NDVI values reflect greater vigor and photosynthetic capacity of the canopy, whereas lower NDVI values are indicative of vegetative stress resulting in chlorophyll reductions and changes in the leaves’ internal structure due to wilting (Gu et al. 2007). NDVI is a commonly used satellite-derived index for monitoring vegetation vigor and tracking drought (e.g., Ji and Peters 2003).

Weekly NDVI data valid for CAR11 during the growing season were provided by the Canadian Crop Condition Assessment Program (CCAP). In brief, a land-use mask is applied in which a pixel is classified as crop/pasture if at least 50% of that pixel is crop/pasture. Here, NDVI data for only those pixels classified as crop/pasture are included in the calculation of NDVI for each CAR. However, it is possible that some pixels that are classified as crop/pasture could have 40% forest cover. About 89% of CAR11 is classified as cropped land or pasture. CCAP uses data from the National Oceanic and Atmospheric Administration satellites carrying the Advanced Very High Resolution Radiometer (AVHRR). A composite of AVHRR images covering a 7-day period removes most or all cloud effects (Latifovic et al. 2005). Corrections are also made to remove much of the atmospheric contamination and to minimize other effects (e.g., view angle and solar angle).

2.4. Atmospheric sounding data

Sounding data are very useful for quantifying and describing two of the criteria required for thunderstorms: the presence of instability and low-level moisture. Additionally, the sounding data can be used to quantify the effects, if any, of changes in vegetation vigor on the structure and characteristics of the convective boundary layer (e.g., Desai et al. 2006).

For this purpose, we examined 0000 UTC weather balloon data from the sounding site at Stony Plain (53.53°N, 114.10°W) from 1 June to 31 August for 1999–2009 (see Figure 1). The 0000 UTC release time falls within the window of maximum thunderstorm activity over central Alberta. During the study period, over 1000 soundings were released from Stony Plain. Before the soundings could be used to calculate sounding parameters, each sounding was visually inspected to determine whether errors or missing data were present and whether the sounding had been contaminated by rain. This resulted in 81 soundings being excluded from our analysis. Soundings released in precipitation (i.e., saturated profile) accounted for the vast majority of contaminated soundings. Soundings for only 36 (less than 10%) of the 365 days when lightning was observed over CAR11 (for the 33 summer months considered) were excluded from the data analysis. Consequently, we are confident that the soundings captured the mean conditions observed during each of the 33 summer months. Because the sounding data are from a specific point in time, they may not always be representative of prestorm conditions over the entire CAR for the 0930–1930 LST (1700–0300 UTC) window. CAR11 has an equivalent radius of just over 70 km, which lies within the range criteria for proximity soundings typically applied in the literature (e.g., Potvin et al. 2010). The temporal criteria for proximity soundings cited in Potvin et al. (Potvin et al. 2010) range from ±30 min from the sounding time to 6 h before and 3 h following the sounding time.

A commercial software package called the Rawinsonde Observation Program (RAOB) was used to batch process the sounding data. These data were then aggregated into monthly and seasonal means for the purpose of calculating standardized anomalies (see section 2.6). RAOB was also used to generate composite soundings (i.e., mean profiles) for each of the 33 months.

2.5. NARR and PAMII data

Although the sounding data were useful for describing the thermodynamic and moisture profiles, they did not offer insight into the vertical motion and moisture advection occurring over CAR11. Here we considered the 500-mb vertical pressure velocity (omega or ω500) and the 0–30-mb above ground moisture flux convergence (MCON) from the North American Regional Reanalysis (NARR; Mesinger et al. 2006). The vertical velocity data were used to quantify the degree of synoptic-scale ascent (or descent) that favors (or inhibits) deep convection. The MCON was used to quantify how much near-surface moisture was being introduced into (or removed from) the study area. NARR data were available on a 0.3° by 0.3° grid. The daily-mean, area-averaged 500-mb omega and MCON data were calculated by averaging the 3-hourly values (for 25 NARR grid points located over CAR11) between 1800 and 0300 UTC. The daily data were then used to calculate monthly means.

Modeled ET data from the Second Generation Prairie Agrometeorological Model (PAMII; see Raddatz 1993) are also presented in the paper. Monthly ET totals were calculated from daily output. Brimelow et al. (Brimelow et al. 2010a; Brimelow et al. 2010b) validated PAMII against in situ observations and found it successfully captured the interannual and intra-annual variability in soil moisture and ET in Alberta.

2.6. Calculation of standardized anomalies and correlations

All monthly and seasonal NDVI, DUR, and sounding data were standardized using the methodology employed by Koster and Suarez (Koster and Suarez 2004). Specifically, the monthly (June, July, and August) and seasonal (June through August) standardized anomalies were calculated as follows:
e1
where Xj is the value of NDVI (or DUR or sounding parameter) for a given month j [or a particular summer (JJA)], is the mean value of X over the 11 years, and sigma is the standard deviation for month j over the 11 years. The weekly NDVI data were used to calculate the monthly and seasonal means. The start and end dates for the seasonal and monthly data varied slightly from year to year, with start dates ranging from 3 to 9 June and end dates ranging from 1 to 7 September. This, however, is not expected to significantly affect the seasonal NDVI anomalies, because there was typically very little change in NDVI values during these times.

The association between NDVI and DUR was quantified using the coefficient of determination (R2) between pairs of standardized anomalies of lightning duration (ΔDUR), NDVI (ΔNDVI), and 500-mb omega (Δω500) and all sounding-derived parameters for individual months and also by season (JJA) for a 11-yr period spanning 1999 through 2009. The statistical significance of the correlations was quantified using p values (between 1% and 10%). It is important to keep in mind that R2 values do not speak to causality between variables. Two-sample t tests and two-sample proportion tests were used to quantify the statistical significance of differences between two population means and proportions, respectively.

2.7. Surface data and upper-air classification

In this section, we use both subjective and objective analyses to quantify the surface and upper-air synoptic-scale circulation patterns (i.e., surface and upper-air forcing) over CAR11 for 22 summer months with contrasting NDVI and/or DUR. By “contrasting,” we mean months with standardized NDVI and DUR anomalies of at least ±0.5. Here, by “upper-air forcing,” we refer to 500-mb vertical motion (ω500) to identify days and months when upper-air conditions were favorable (or not) for thunderstorm activity. The rationale for identifying this subsample was to closer examine the role of NDVI and synoptic-scale forcing in modulating DUR.

The relationship between tropospheric ascent and thunderstorm activity in the midlatitudes is well established, and this must be accounted for in our analysis. Doswell (Doswell 1987), Moller (Moller 2001), and Schumann and Roebber (Schumann and Roebber 2010) note that synoptic-scale disturbances act to destabilize the atmosphere and weaken the capping lid through slow ascent, which favors the formation of widespread convection. Synoptic-scale systems (in particular upper-air troughs) have long been recognized as a factor in modulating thunderstorm activity over central and southern Alberta (e.g., Strong 1986; Smith and Yau 1993a; Smith and Yau 1993b).

Given the oftentimes rapid evolution of the upper-air field and the paucity of upper-air analyses (i.e., only available at 1200 and 0000 UTC), we objectively quantified whether the synoptic-scale upper-air environment was favorable for thunderstorm activity by calculating the mean, area-averaged 500-mb daily omega from the NARR as described in section 2.5. We applied a threshold of −1.0 μbar s−1 to identify days with organized and persistent ascent. This value is lower than that cited in the literature. For example, Wang et al. (Wang et al. 2009) used a threshold of −2 μbar s−1 for 600-mb velocity to track midtropospheric perturbations. Stensrud and Fritsch (Stensrud and Fritsch 1993) considered events associated with omega greater than −1 μbar s−1 to be “weakly forced.” The difference in thresholds between this study and others is not surprising, given that the value used here represents the mean of 25 grid points over a 9-h period, whereas others have typically discussed instantaneous values at point locations.

Raddatz and Hanesiak (Raddatz and Hanesiak 2008) investigated almost 1000 significant summer rainfall events (≥10 mm in 24 h) on the Canadian Prairies from 2000 to 2004. They found that surface features play an important role in initiating and organizing thunderstorms in this region. In this study, we adopted a similar method to that of Raddatz and Hanesiak (Raddatz and Hanesiak 2008) to identify robust surface features (e.g., low and high pressure systems, fronts, and troughs). Specifically, 3-hourly maps between 1800 and 0300 UTC of mass divergence, equivalent potential temperature, streamlines, and surface pressure were generated using the Plymouth State University Weather Center interactive plotting tool. The maps were generated using all available surface observations and then subjectively examined to identify robust surface features. With the exception of surface lows and highs (the centers of which had to occur within 200 km of CAR11; see Raddatz and Hanesiak 2008), all other features had to have moved across or been situated over CAR11 at some point between 1800 and 0300 UTC.

The surface and midtropospheric ascent data described above were then combined for each day and assigned one the following six classes using the decision tree shown in Figure 2: high pressure or ridge (HP); surface forcing only; upper-air forcing only; surface or upper-air forcing; surface and upper-air forcing, and mesoscale. The mesoscale class is defined as a day with no robust surface features and with mean 500-mb omega greater than −1.0 μbar s−1.

Figure 2.
Figure 2.

Decision tree used to classify the surface and upper-air pattern on each day. See text for details. HP refers to a cell of high pressure.

Citation: Earth Interactions 15, 31; 10.1175/2011EI407.1

3. Response of lightning activity to synoptic-scale forcing

Brimelow et al. (Brimelow et al. 2011) suggested that a positive feedback mechanism was evident during the summer months over drought-affected regions on the Canadian Prairies, with drought conditions leading to less lightning activity. They also noted that the relationship between NDVI and DUR increased significantly as both the area and magnitude of the dry anomaly increased. For CAR11 for the 33 summers months from 1999 through 2009, a statistically significant relationship was found between NDVI and DUR (R2 = 0.14 and p = 0.03). This R2, albeit statistically significant, is relatively low, which suggests that other factors are at play in modulating DUR. Before discussing the role of vegetation vigor in modulating lightning activity over CAR11, we will first explore the relationship between synoptic-scale forcing and lightning activity.

3.1. Relationship between 500-mb omega and DUR

As a first step toward elucidating the relationship between mean daily 500-mb vertical motion over CAR11 and the concomitant DUR, we examined the relationship between ω500 and DUR by partitioning the velocity for all days into two groups: ω500 < 0.0 μbar s−1 (mean ascent) and ω500 > 0.0 μbar s−1 (mean descent). Here, 50% of the days with mean ascent over CAR11 produced lightning. In contrast, only 23% of the days with mean descent over CAR11 produced lightning. Further partitioning of the lightning data according to ω500 revealed that 62% of days with marked upper-air ascent (ω500 < −1.0 μbar s−1) produced thunderstorms over CAR11, whereas 32% of days with intermediate vertical motion (−1.0 μbar s−1 < ω500 < +1.0 μbar s−1) produced lightning. In contrast, only 16% with marked descent (ω500 > +1.0 μbar s−1) produced lightning. Thus, a systematic increase in the likelihood of storms is evident as vertical ascent increases.

Figure 3 shows that thunderstorm days occurring when there was marked midlevel ascent over CAR11 tended to be associated with more lightning activity [mean DUR of 3.0 h, with 95% confidence interval (CI) between 2.7 and 3.4 h] than on days when storms developed when there was marked midlevel descent (mean DUR of 1.5 h, with 95% CI between 0.7 and 2.2 h). The mean DUR values for these two contrasting ω500 classes are different at the 99% confidence level.

Figure 3.
Figure 3.

Duration over CAR11 for days with contrasting mean area-averaged ω500 (omega) and when thunderstorms were observed. Crosshairs represent the mean.

Citation: Earth Interactions 15, 31; 10.1175/2011EI407.1

The R2 between monthly ΔDUR and Δω500 for 33 summer months was 0.187 (p = 0.012). The relationship was especially marked in July (R2 = 0.44 and p = 0.03) and June (R2 = 0.30 and p = 0.08) but virtually nonexistent in August (R2 = 0.01 and p = 0.806). Hence, ω500 explains a significant amount of variance of DUR for the 33 summer months considered, especially in June and July combined (R2 = 0.37, p = 0.003, and N = 22). An investigation of the apparent breakdown in R2 between ΔDUR and Δω500 in August is beyond the scope of this study.

3.2. Relationship between both surface and upper-air forcing and DUR

In section 3.1, we considered only midtropospheric vertical motion. However, surface features also play an important role in triggering and organizing thunderstorms (e.g., Raddatz and Hanesiak 2008). Consequently, in order to further explore the relationship between DUR and synoptic-scale forcing, we studied the linkages between monthly DUR anomalies associated with the six forcing classes for the subset of 22 months referred to in section 2.7. The months included in this subset and values for selected parameters are listed in Table 1.

Table 1.

Summary of 22 summer months between 1999 and 2009 with contrasting ΔNDVI and/or ΔDUR. “Contrasting” means months with standardized NDVI and DUR anomalies of at least ±0.5 (with weekly standardized NDVI anomalies for at least 15 days exceeding ±0.5). Other standardized monthly anomalies are for ω500 = NARR 500-mb vertical velocity; ΔMCON = NARR 0–30-mb above-ground moisture flux convergence; height of the LCL from Stony Plain soundings; mixing height (MH) from soundings; r50 = mean mixing ratio in lowest 50 mb above ground from soundings; and Γ850–700 = lapse rate between 850 and 700 mb. See text for more details. Here, negative standardized anomalies for ω500, MCON, and Γ850–700 represent above-average ascent, weaker than average MCON, and a stronger lapse rate between 850 and 700 mb, respectively.

Table 1.

Three metrics were then calculated for each forcing class and for both NDVI groups. The first metric is the mean DUR calculated for all days in the group, including days with no lightning. The second metric is the success rate, which refers to the percentage of days in a given class that produced lightning over CAR11. The third metric is the mean DUR for only those days in each class when lightning was observed.

Figure 4 shows that the lowest mean DUR was observed for days in the HP class (N = 124), with the longest mean DUR observed for days in the surface and upper-air class (N = 60). That is, the most lightning was typically observed on days when a robust surface feature was present in tandem with midtropospheric ascent. Two-sample t tests indicated that the difference in mean DUR between adjacent classes was statistically significant (90% level) but not between the surface only (N = 74) and upper-air only (N = 64) classes.

Figure 4.
Figure 4.

DUR (solid green line), success rate (dashed red line), and DUR on storm days (solid blue line) for various forcing classes identified using the decision tree in Figure 2. Error bars for the DUR data represent the 95% CI of the class mean.

Citation: Earth Interactions 15, 31; 10.1175/2011EI407.1

A similar pattern is evident in Figure 4 for the mean DUR on storm days, with DUR increasing from 1.4 h for the HP class (N = 13) to 3.5 h for the surface and upper-air class (N = 43). The differences in mean DUR were statistically significant (90% confidence level) only between the surface or upper-air (N = 113) and the surface and upper-air classes. The difference in DUR for storm days was also statistically significant between the surface and upper-air and the upper-air only (N = 30) classes (99% confidence level). The difference in mean DUR on storm days between mesoscale (N = 48) and HP, between upper-air only and surface only (N = 40), and between upper-air or surface and upper-air only classes was statistically different at only the 80% level of confidence. The smaller sample sizes in each class for storm days precluded distinguishing between class means at statistically significant levels.

The success rate increases as one progresses from the HP class (about 10%) to the surface and upper-air class (72%) (Figure 4). Proportion tests between adjacent classes determined that the differences in success rate were statistically significant at least at the 90% confidence level. There was one exception: days classified as surface only did have a higher success rate than days classified as upper-air only, but this difference was statistically significant at only the 80% confidence level. These results suggest that surface features are more likely to initiate and extend lightning activity than midtropospheric ascent alone and that surface and upper-air features acting in concert are especially effective at initiating and organizing thunderstorms.

3.3. Isolating the impact of NDVI on DUR

In previous sections, we demonstrated that synoptic-scale forcing notably affects lightning activity over CAR11, and any potential signal from NDVI could be swamped by strong synoptic forcing. Consequently, extracting a signal from the contribution (if any) of vegetation vigor to DUR is difficult.

To address this issue, we looked at the surface and upper-air forcing for the subset of 22 months with contrasting DUR and NDVI. Specifically, we considered the response of DUR on days with similar synoptic forcing but occurring during weeks with contrasting NDVI conditions (+0.5 ≤ NDVI ≤ −0.5). For example, all days classified as mesoscale were cross-referenced with the weekly NDVI anomalies for CAR11. If a day was classified as mesoscale and occurred during a week when NDVI was ≤−0.5, it was grouped into the much-below-average (MBA) class. Similarly, those days classified as mesoscale that were observed during weeks when NDVI was ≥+0.5 were added to the much-above-average (MAA) class (Table 2). In this way, the effect of synoptic forcing was minimized thereby allowing us to focus on the response, if any, of the lightning to changes in antecedent vegetation vigor.

Table 2.

Lightning metrics by forcing class for contrasting NDVI conditions. MAA refers to much above average NDVI (ΔNDVI > +0.5), and MBA refers to much above average NDVI (ΔNDVI < −0.5). Sample size for each class is denoted by N.

Table 2.

As was observed in section 3.2, a systematic increase in mean DUR, success rate, and mean DUR on storm days is seen as the forcing becomes more organized (Table 2). This was true regardless of the vegetation vigor. Differences for mean DUR for contrasting vegetation vigor were statistically different for days classified as mesoscale (p = 0.042 or >95% confidence) and as surface or upper air (p = 0.171 or >80% confidence). The differences were more marked when considering the mean DUR on storm days only, with statistically significant differences obtained for the mesoscale (p = 0.018 or >95% confidence), upper-air only (p = 0.107), and surface or upper-air (p = 0.110 and almost 90% confidence) classes. The success rate when NDVI was much above average was statistically greater than when NDVI was much below average for only the surface and upper-air class.

This analysis suggests that longer DUR can be expected when storms develop on days when NDVI is much above average than on days having similar synoptic-scale forcing but with much-below-average NDVI. Also, on days when a surface or upper-air forcing is present, the occurrence of lightning is higher (and is of longer duration) when NDVI over CAR11 is much above average, compared to when NDVI is much below average.

3.4. Overview of impact on forcing and NDVI on DUR

The interplay between midtropospheric omega and NDVI anomalies on DUR over CAR11 for 33 summer months from 1999 through 2009 can be visualized using a surface map as shown in Figure 5. The parameter space may be roughly divided into three zones as indicated by the dashed lines:

  • Zone I: Below-average ΔDUR (<−0.5) is typically observed when ΔNDVI < −0.5, even when the midtropospheric ascent is above average (Δω500 < 0.0).

  • Zone II: If 500-mb omega is relatively weak (Δω500 > +0.5), then ΔDUR is unlikely to be above average even if ΔNDVI is much above average.

  • Zone III: If ΔNDVI is above average and 500-mb omega is favorable (Δω500 < +0.5), then ΔDUR is highly likely to be above average.

Figure 5.
Figure 5.

Monthly ΔDUR (shading) as a function of ΔNDVI and Δω500 (standardized monthly 500-mb omega anomalies). Black dots represent individual data points for each of the 33 summer months. Dashed lines delineate the three zones discussed in the text. Data were interpolated using a kriging scheme.

Citation: Earth Interactions 15, 31; 10.1175/2011EI407.1

A couple of months in Figure 5 are clearly outliers, but an inspection of data from Table 1 sheds some light as to why this might be. It should be noted that, for the 33 months considered here, MCON was poorly correlated with DUR and with low-level moisture and is barely statistically significantly correlated with monthly NDVI. Nevertheless, MCON does explain the discrepancies for some months. For example, June 2001 had near-normal omega at 500 mb and below-average NDVI, but DUR was much above average. This may, in part, have been attributable to the strong MCON, which, despite the stressed vegetation, led to near-normal low-level mixing ratios. The resultant lower cloud-base heights, in combination with near-normal ascent, may explain the above-average DUR in June 2001.

4. Relationship between sounding data, NDVI, and DUR

In this section, we explore relationships between monthly standardized anomalies of sounding parameters derived from 0000 UTC Stony Plain soundings and the concomitant standardized anomalies of NDVI and DUR over CAR11. We will focus on eight sounding-derived variables that are significantly correlated with both NDVI and DUR. The rationale is to investigate the mechanisms by which anomalies in NDVI could affect the sounding structure in the low levels, which could in turn affect thunderstorm activity.

Of the parameters summarized in Table 3, three (Td850, r50, and Tdsfc) represent low-level moisture, two [Γ850–700 and mixing height (MH)] represent low-level lapse rates, two [LCL and convective condensation level (CCL)] represent the integrated effects of low-level moisture and temperature, and one (Tcon) represents the effects of low-level moisture and low-level lapse rates. Table 3 indicates that, for all Junes, Julys, and Augusts combined, the highest R2 values were between ΔLCL (and ΔMH) and both ΔNDVI and ΔDUR. The weakest correlations were for ΔTd850 and ΔTcon versus ΔNDVI and ΔDUR. Table 4 shows that, in August, only low-level moisture variables (Td850, r50, and Tdsfc) were positively correlated with DUR at statistically significant levels.

Table 3.

Comparison of R2 between standardized anomalies of selected sounding variables (from 0000 UTC Stony Plain soundings) and standardized anomalies of NDVI and DUR over CAR11 in Alberta for different subsets of data. Shown are all Junes and Julys combined (JJ); all Julys and Augusts combined (JA); all Junes, Julys, and Augusts (33 months) combined (JJA); and July through August for each of the 11 years (JJA). Also shown are CCL height, mixing height; surface dewpoint Tdsfc, dewpoint at 850 mb Td850; convective temperature Tcon; mean mixing ratio at 50 mb above ground level r50; and lapse rate between 850 and 700 mb Γ850–700. Columns under the header “correlation” indicate the sign of the correlation between variables.

Table 3.
Table 4.

As in Table 3, but for individual months.

Table 4.

For 33 summer months at Stony Plain, ΔLCL and ΔNDVI were negatively correlated, with LCL heights decreasing as NDVI increased (Figure 6). The relationship between ΔLCL and ΔDUR was also negative, with DUR increasing as the LCL heights decreased. This is consistent with the findings made by Peppler and Lamb (Peppler and Lamb 1989), who found that area-averaged rainfall and LCL heights calculated from Stony Plain sounding data were negatively correlated for May through August for the three years considered in their study.

Figure 6.
Figure 6.

Scatterplots of ΔNDVI vs ΔLCL and between ΔLCL and ΔDUR for 33 summer months from 1999 to 2009 over CAR11.

Citation: Earth Interactions 15, 31; 10.1175/2011EI407.1

The above relationship between ΔNDVI and ΔLCL is intuitive, because ΔNDVI is also positively correlated with low-level moisture (Table 3). This suggests that the increase in LCL heights with decreasing vegetation vigor is occurring primarily in response to a decrease in low-level moisture, which is in turn related to decreased ET from drier soils and sparser vegetation. This is supported by the statistically significant relationship between ΔNDVI over CAR11 and standardized anomalies of monthly accumulated ET at Edmonton International Airport (~30 km southeast of Stony Plain) from 1999 to 2005. For the period 1999 through 2005, the R2 between ΔNDVI and ΔET was 0.334 (p = 0.006 and N = 21), which indicates that ΔNDVI and ΔET are closely related. A statistically significant correlation (R2 = 0.58 and p = 0.000) was also found between ΔNDVI and ΔET for summers between 1999 and 2004 using data for 12 CARs on the Canadian prairies having equivalent radii of 75 km more (data provided by R. Raddatz 2009, personal communication).

All three low-level moisture variables are positively correlated with ΔNDVI, which is consistent with the relationship between NDVI and ET discussed above. Namely, decreased vegetation vigor is associated with lower ET (i.e., latent heat flux) into the CBL. The moisture variables are in turn all positively correlated with ΔDUR, with a decrease in low-level moisture typically associated with a decrease in thunderstorm activity (i.e., DUR). This is expected, given that low-level moisture is a necessary condition for thunderstorms.

The relationship between ΔMH and ΔNDVI is negative, suggesting that, as vegetation vigor decreases, MH increases (i.e., deeper CBL). MH is also negatively correlated with DUR, suggesting that a deeper CBL over this region is typically unfavorable for thunderstorm activity. MH is negatively correlated with low-level moisture parameters (not shown), which suggests that low vegetation vigor results in a deeper, drier CBL on account of a higher sensible heat flux. The relationships between ΔNDVI and both ΔCCL height and Tcon are also negative. This reflects the higher Tcon and increasing CCL height with decreasing vegetation vigor. These particular changes in Tcon and CCL height are likely being modulated by decreases in low-level moisture associated with reduced vegetation vigor and concomitant ET. CCL height and Tcon are also both negatively correlated with DUR. Specifically, as Tcon increases, DUR typically decreases. This is not surprising as higher convective temperatures make it increasingly unlikely that solar heating alone will provide sufficient lift to initiate storms.

Here, Γ850–700 is positively correlated with both ΔNDVI and ΔDUR, with lapse rates becoming more negative (i.e., stronger) as vegetation vigor decreases. This is consistent with a deeper and warmer CBL observed on account of increased sensible heat flux associated with below-average NDVI. The positive correlation between lapse rate and DUR suggests that, in this region, as the 850–700-mb lapse rate becomes weaker, DUR increases. Although weaker lapse rates being associated with greater DUR may seem counterintuitive, this could suggest that, on average, a shallower (but more moist) CBL is more likely to support thunderstorms than is a deeper, drier CBL.

Here, we discuss the composite soundings for two contrasting Junes. In 2002, drought conditions were present over CAR11 (ΔNDVI = −1.6), with much-below-average lightning duration (ΔDUR = −0.9). In 2005, pluvial conditions were observed over CAR11 (ΔNDVI = +1.6), with much-above-average duration (ΔDUR = +0.8). The composite soundings in Figure 7 show that the LCL height for the 2002 composite sounding was much higher at 2783 m above ground level (at 2°C), compared to 1974 m (at 6°C) in 2005. The mean mixing ratio in the lowest 50 mb in 2002 was 5.5 g kg−1, compared to 6.6 g kg−1 in 2005. The mean root-zone plant-available soil moisture content in 2005 (as predicted by PAMII) for Edmonton International Airport was 71%, with a predicted accumulated monthly ET of 54 mm. In contrast, in 2002, the mean root-zone soil moisture was 32%, with a predicted accumulated monthly ET of 41 mm (about 25% less than in 2005).

Figure 7.
Figure 7.

Composite soundings for Stony Plain in June 2002 (red lines), when CAR11 was experiencing drought conditions, and in June 2005 (blue lines), when CAR11 was experiencing pluvial conditions.

Citation: Earth Interactions 15, 31; 10.1175/2011EI407.1

It is possible that (in contrast to 2002) the upslope flow in 2005 was partly responsible for increasing the depth of moisture in the CBL over CAR11 (Figure 7). However, the monthly means of 0–30 mb (above ground) moisture flux convergence (from NARR) over CAR11 for June 2002 and for June 2005 were very similar at −1.49 × 10−8 s−1 and −1.81 × 10−8 s−1, respectively (Table 1). Therefore, the NARR data indicate that moisture flux divergence of similar magnitude was present over CAR11 for both months. Another clue that differences in CBL moisture in Figure 7 were not primarily the result of advection is that the moisture traces only diverge within the CBL (below about 800 mb), with the greatest differences being observed at the surface. It is possible LCL heights could be changing in response to large-scale forcing without invoking a connection to the surface. That is plausible, but only if the LCL heights are being modulated by large-scale forcing and/or large-scale moisture advection. Inspection of the data did not reveal compelling evidence for this hypothesis, with no relationship evident between MCON and LCL heights and a relatively weak relationship between ω500 and LCL heights (R2 = 0.106 and p = 0.065).

Hence, moisture advection is unlikely the primary reason for the differences in low-level moisture. Rather, ET data (from PAMII) suggest that reduced latent heat flux (and higher sensible heat flux) on account of the low soil moisture and stressed vegetation were largely responsible for the deeper, warmer, and drier boundary layer in 2002. This claim is consistent with the differences in key NARR fields (not shown) such as soil moisture, latent and sensible heat fluxes, and boundary layer depth for contrasting months.

All things being equal (e.g., similar synoptic-scale forcing), the drier and warmer CBL in June 2002 would have limited thunderstorm activity compared to June 2005. In fact, considering days with surface or upper-air forcing, the mean-daily DUR for storm days in June 2002 was 2.1 h compared to 2.8 h in 2005. The success rate in 2005 was 73% compared to 60% in 2002.

Brimelow et al. (Brimelow et al. 2011) indicated an asymmetric relationship between ΔNDVI and ΔDUR for CARs on the Canadian Prairies, with positive ΔNDVI not a necessary or a sufficient condition for above-average lightning duration. Our results here corroborate this, with months with below-average NDVI typically associated with below-average DUR, whereas the opposite was not necessarily true when NDVI was above average. Specifically, R2 between ΔNDVI and ΔDUR for months with much-below-average NDVI was 0.306 (p = 0.12 and N = 9), whereas R2 between ΔNDVI and ΔDUR for months having much-above-average NDVI was 0.001 (p = 0.77 and N = 12).

To investigate the asymmetric response of DUR to NDVI, R2 was calculated between standardized anomalies of selected sounding variables and standardized anomalies of NDVI and DUR for months having below-average NDVI and above-average NDVI. Table 5 summarizes the correlations for months grouped according to whether NDVI was greater or less than 0.0; this threshold was applied to increase the sample size. For below-average ΔNDVI, both ΔCCL and ΔLCL were correlated with ΔNDVI and ΔDUR at the 90% confidence level, whereas ΔTdsfc was correlated with ΔNDVI at the 90% confidence level but was not so with ΔDUR. Here, ΔTcon was statistically significantly correlated with ΔNDVI and with ΔDUR at marginally statistically significant levels (p = 0.115). In contrast, when NDVI was above average, only ΔLCL and ΔMH were correlated with DUR at the 90% confidence level. Thus, the coupling between the surface condition, sounding parameters and lightning activity is not robust when NDVI is above average.

Table 5.

Comparison of R2 between standardized anomalies of selected sounding variables and standardized anomalies of NDVI and DUR over CAR11 in Alberta for months having below-average NDVI (ΔNDVI < 0.0) and above-average NDVI (ΔNDVI > 0.0).

Table 5.

Regression models

Here, we investigate the impact of using different combinations of selected sounding parameters and midtropospheric omega on the variance explained in DUR. The coefficients of determination for multiple linear regression models for all 33 months are shown in Table 6. Results are also shown for the subset of 22 months to assess the impact of including information about the surface forcing. Because ΔLCL and ΔΓ850–700, ΔLCL and ΔMH, and ΔMH and ΔΓ850–700 were strongly correlated, regression models were run using only one of the variables from each pair to avoid introducing redundancy. The R2 between the number of days per month with surface or upper-air (SUA) forcing (∑SUA) and ω500 (<−1 μbar s−1) was 0.64. NDVI was not correlated with ω500 or ∑SUA. Together, ΔLCL (which implicitly includes low-level moisture) and Δω500 explain most of the variance in DUR, which underscores the importance of synoptic lift and moisture for thunderstorm activity. The variance in ΔDUR explained using ΔNDVI and Δω500 is virtually the same as that explained using ΔTdsfc and Δω500 or using Δr50 and Δω500 (not shown). These results suggest that NDVI is a suitable proxy for low-level moisture. The relationships between sounding variables and DUR for the subset of 22 months were very similar to those for all 33 months. The most noticeable difference is that including information about surface features increases the variance in DUR explained over using 500-mb omega alone.

Table 6.

Coefficients of determination R2 for various multiple linear regression models. When more than one predictor was used, the adjusted R2 (R2adj) is provided. Values in parentheses are p values. Plus symbols indicate which variables (i.e., predictors) were used in the correlation. The bottom half of the table indicates which variables were used for the subset of 22 months.

Table 6.

5. Conceptual model and discussion

The relationships between synoptic-scale forcing, NDVI, CBL characteristics, and DUR discussed in previous sections can be summarized in the conceptual model presented in Figure 8. In our conceptual model, we assume that large-scale forcing initially results in a significant precipitation deficit that in turn results in reduced vegetation vigor. Liu et al. (Liu et al. 2004) studied the moisture transport and upper-air circulation patterns for the 2000–01 drought over the Canadian Prairies. They noted that, during summer drought, moisture transport from the Gulf of Mexico was reduced on account of a persistent upper-air blocking and that attendant large-scale subsidence lowered precipitation. The associated changes in vegetation (i.e., NDVI) in response to soil moisture anomalies can have a marked impact on the partitioning of available energy (e.g., Liu et al. 2006), which in turn affects the structure (and moisture content) of the CBL (e.g., Taylor et al. 2010) and potentially thunderstorm activity.

Figure 8.
Figure 8.

Conceptual model describing the relationships between NDVI, synoptic-scale forcing, CBL characteristics, and DUR over CAR11. Red crosshair symbols represent a positive correlation, and blue symbols represent a negative correlation.

Citation: Earth Interactions 15, 31; 10.1175/2011EI407.1

As shown in Figure 8, our analysis of sounding, NDVI, and lightning data indicates that a reduction in vegetation vigor—in response to soil moisture deficits—modulates the partitioning of available energy into sensible and latent heat fluxes at the surface. Under such conditions, the latent heat flux is reduced while the sensible heat flux is increased, thereby causing higher surface temperatures and lower moisture content in the CBL. These factors combined result in a warmer, drier, and deeper CBL, with steeper low-level lapse rates. Additionally, the higher temperatures and lower near-surface moisture increase the height of the LCL. This acts to lower moist static energy, lower the in-cloud liquid water content of clouds, and also increase the evaporation of any precipitation below cloud base (e.g., Yamada 2008). Together, these factors result in a positive feedback between vegetation vigor and DUR. We have demonstrated that ΔNDVI is positively correlated with low-level moisture, which suggests that the increase in LCL heights with decreasing vegetation vigor is occurring primarily in response to a decrease in low-level moisture, which is in turn related to decreased evapotranspiration from drier soils and sparser vegetation. It should be noted that the CCL heights were also strongly correlated with NDVI. For this reason, the CCL height could potentially be substituted for LCL heights in the conceptual model. In the end, we decided upon LCL because it takes into account the impact of both low-level moisture and temperature on cloud-base height.

Our conceptual model also includes the important contribution of synoptic-scale forcing to thunderstorm activity. In fact, the NDVI–DUR coupling is probably of secondary importance to synoptic-scale forcing, at least for this region. As illustrated by the analysis for a subset of 22 months (section 4), including information about surface features in addition to midtropospheric omega significantly increases the variance of DUR explained.

Our conceptual model is consistent with previous research demonstrating that the amount and depth of moisture in the boundary layer is critical in determining the amount of convective instability and thunderstorm intensity. Crook (Crook 1996) demonstrated that small changes in low-level mixing ratio (~1 g kg−1) can have significant effects on the initiation and intensity of thunderstorms. Siqueira et al. (Siqueira et al. 2009) found that the CBL grew more quickly under water-stressed conditions in their model simulations than when soil moisture was plentiful. However, the rapid growth of the CBL did not translate into the triggering of convection because the height of the LCL also increased rapidly under drought conditions. Cloud model simulations of convection by Yamada (Yamada 2008) found that the higher cloud-base height for drier soil conditions produced a lower rain rate because of increased evaporation of precipitation within the deeper and drier subcloud layer. Matyas and Carleton (Matyas and Carleton 2010) found a positive correlation between the intensity of convection and soil moisture over the eastern Corn Belt. Juang et al. (Juang et al. 2007) found that convective precipitation over moist soils was typically greater in intensity and led to higher precipitation accumulation than for events over dry soil conditions. Garrity et al. (Garrity et al. 2010) studied the impact of extreme wet and dry regions over the western United States on atmospheric moisture profiles from reanalysis data. They found that the decrease of moisture during drought conditions was limited to below 700 mb and suggest that this may indicate that thunderstorms are playing a role in modulating droughts.

Our observations over CAR11 suggest that NDVI appears to be modulating DUR by altering the moisture content of the CBL and in turn the LCL height. This corroborates research by Betts (Betts 2004) and Betts and Viterbo (Betts and Viterbo 2005), who found a strong link between a soil moisture index and the height of the LCL in the European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis data. Similarly, Betts (Betts 2007) examined reanalysis data over the Mississippi and found that increased ET increased low-level relative humidity, which in turn lowered the LCL. Betts (Betts 2007) noted that LCL decreased as soil moisture and precipitation increased and that precipitation increased with a lowering of the LCL. Jiang et al. (Jiang et al. 2009) found a similar relationship between LCL, soil moisture, and subsequent precipitation in the Weather Research and Forecasting model and the NARR data.

5.1. Comparison with other conceptual models

It is worth noting that our conceptual model is based largely on observations, unlike other conceptual models proposed for land–atmosphere feedbacks that are typically based on numerical modeling studies (e.g., Kim and Wang 2007; Lawrence and Slingo 2005; Trier et al. 2004; Pal and Eltahir 2001; Schär et al. 1999). With that said, the in situ observational datasets used in this study corroborate the conceptual models based on numerical model simulations. Our conceptual model lends support to the so-called “indirect soil moisture–precipitation feedback” mechanism postulated by Schär et al. (Schär et al. 1999), which describes how changes in soil moisture affect precipitation by influencing CBL characteristics and instability.

Our findings are, however, not consistent with those of Myoung and Nielsen-Gammon (Myoung and Nielsen-Gammon 2010a), who used NCEP–NCAR reanalysis data to investigate mechanisms controlling the variability of convective precipitation. They found that the convective inhibition (CIN) is negatively correlated with precipitation over the continents in July, including over the Canadian Prairies (see their Figure 6c). Myoung and Nielsen-Gammon (Myoung and Nielsen-Gammon 2010b) examined the relationship between soil moisture and precipitation over Texas. They found that reduced precipitation decreases the soil moisture, which increases the surface temperature and lowers the dewpoint, which, together with warmer temperatures at 700 mb, increases CIN. They concluded that CIN is the primary factor maintaining drought over Texas. In contrast, the link between soil moisture and thunderstorm activity in our study appears to be manifested through the LCL height.

5.2. Asymmetry of the land–atmosphere feedback

Brimelow et al. (Brimelow et al. 2011) found an asymmetry between NDVI and DUR for 38 CARs across the prairies, with much-above-average NDVI not a necessary or sufficient condition for above-average lightning duration. Similarly, we also found an asymmetry in the coupling between NDVI and DUR over our study area. Specifically, the coupling between NDVI and DUR is strong when vegetation is less vigorous, with much-below-average NDVI typically associated with higher LCL heights and below-average DUR. However, the opposite is not necessarily true when NDVI is much above average.

We purport that this asymmetry is attributable to the fact that, if moisture in the CBL drops below a critical point, because of reduced transpiration on account of dry soils, then, regardless of the amount of instability and lift, convection is unlikely. In contrast, increasing low-level moisture beyond a critical value means that moisture is no longer the limiting condition for thunderstorm formation; lift and instability then are. This hypothesis is supported by observational studies that have shown the relationship between soil moisture stress and canopy conductance is asymmetrical and can be described by a logistic regression model (e.g., Anderson et al. 2007). In particular, observations find that there is little impact on canopy conductance (i.e., transpiration) for root-zone plant available water content as low as 50%, but a rapid decrease in canopy conductance and photosynthetic activity is observed below that point (e.g., Schulze 1986; Mitchell et al. 2001). Although the asymmetric relationship between NDVI (or soil moisture) and precipitation has been identified in several numerical modeling studies (e.g., Pal and Eltahir 2001; Oglesby et al. 2002; Hohenegger et al. 2009), our research here and in Brimelow et al. (Brimelow et al. 2011) is to the best of our knowledge the first body of work to confirm the existence of this asymmetric feedback using largely observed data.

6. Conclusions

This study focuses on how anomalies in soil moisture and vegetation (i.e., NDVI anomalies) over the Canadian Prairies can condition the convective boundary layer (CBL) so as to inhibit or facilitate thunderstorm activity while also considering the role of synoptic-scale forcing on modulating summer thunderstorm activity. Our study focused on a census agricultural region over central Alberta (CAR11) for which we had observed lightning data, NDVI data, and in situ sounding data for 11 summers (1 June through 31 August; 1999–2009). We used lightning duration (DUR) as a proxy for thunderstorm occurrence and persistence.

To identify whether vegetation anomalies had any impact on lightning duration over CAR11 for similar synoptic-scale forcing conditions, we adopted an ingredients-based approach to identify the three criteria required for thunderstorm formation: instability, low-level moisture, and a trigger mechanism. Sounding data from Stony Plain were examined for the purpose of quantifying and describing the presence of instability and low-level moisture and to quantify the impacts of changes in vegetation vigor on the structure of the CBL. We used both subjective (surface analysis) and objective analyses (NARR 500-mb omega) to quantify the surface and upper-air synoptic-scale forcing each day for 22 summer months with contrasting NDVI and/or DUR. Our primary findings pertaining to CAR11 may be summarized as follows:

  • Lightning duration over CAR11 is greater when storms develop in an environment with favorable synoptic-scale conditions. Also, the amount of variance in monthly DUR explained increases significantly when incorporating information on both surface features and upper-air forcing.

  • The highest lightning occurrence and the longest DUR were typically observed on days when surface forcing and upper-air forcing were synchronized.

  • Longer DUR can be expected when storms develop over areas with high vegetation vigor, regardless of the degree of synoptic-scale forcing.

  • Below-average DUR was typically observed when NDVI was below average, even when the midtropospheric ascent was above average. If the midtropospheric ascent was relatively weak, then DUR was unlikely to be above average, even if NDVI was much above average.

  • Statistical analysis of selected sounding parameters (e.g., LCL height) and NDVI suggests that the vegetation is affecting the structure and evolution of CBL, which in turn is affecting DUR.

  • Analysis of the sounding data in concert with the NDVI and DUR data suggests that the response of DUR to anomalies in NDVI is marked for low vegetation vigor but is not necessarily discernible when NDVI is above average.

We propose a conceptual model that describes how a reduction of vegetation vigor—in response to soil moisture deficits—modulates the partitioning of available energy into sensible and latent heat fluxes at the surface. We also demonstrate that NDVI is positively correlated with low-level moisture, which suggests that the increase in LCL heights associated with decreasing vegetation vigor is occurring primarily in response to a decrease in low-level moisture, which is in turn related to decreased ET from drier soils and sparser vegetation. Our proposed conceptual model is consistent with similar conceptual models but is unique in that it is the first framework based largely on observational data (rather than results from numerical modeling experiments or reanalysis data) to describe the coupling between vegetation, boundary layer characteristics, synoptic-scale forcing, and lightning activity. Our findings lend strong support to the hypotheses that the coupling between vegetation vigor and antecedent lightning duration appears to be manifested through LCL heights and that the response of thunderstorm duration to NDVI anomalies is asymmetric.

Given that the memory of NDVI is on the order of weeks to several months, our work may be of aide in the prediction of thunderstorm duration and occurrence on the monthly time scale. Finally, a caveat: although we have endeavored to isolate the impact of NDVI on DUR, it is possible that synoptic-scale processes are also modulating LCL heights. Future work could focus on shorter time windows (7–14 days), thereby permitting lead–lag correlation analyses between key parameters to confirm that NDVI does in fact lead DUR.

Acknowledgments

This work has been financially supported by a grant to Dr. John Hanesiak from the Drought Research Initiative (DRI), a Canadian Foundation for Climate and Atmosphere Sciences (CFCAS) network. We are also very grateful to Mr. Patrice Constanza and Dr. George Liu who were invaluable in preparing and processing the North American Regional Reanalysis data used in this research.

References

  • Alfieri, J. G., P. Claps, P. D’Odorico, F. Laio, and T. M. Over, 2008: An analysis of the soil moisture feedback on convective and stratiform precipitation. J. Hydrometeor., 9, 280291.

    • Search Google Scholar
    • Export Citation
  • Anderson, M. C., J. M. Norman, J. R. Mecikalski, J. A. Otkin, and W. P. Kustas, 2007: A climatological study of evapotranspiration and moisture stress across the continental United States based on thermal remote sensing: 1. Model formulation. J. Geophys. Res., 112, D10117, doi:10.1029/2006JD007506.

    • Search Google Scholar
    • Export Citation
  • Beljaars, A. C. M., P. Viterbo, M. J. Miller, and A. K. Betts, 1996: The anomalous rainfall over the United States during July 1993: Sensitivity to land surface parameterization and soil moisture anomalies. Mon. Wea. Rev., 124, 362383.

    • Search Google Scholar
    • Export Citation
  • Betts, A. K., 2004: Understanding hydrometeorology using global models. Bull. Amer. Meteor. Soc., 85, 16731688.

  • Betts, A. K., 2007: Coupling of water vapor convergence, clouds, precipitation, and land surface processes. J. Geophys. Res., 112, D10108, doi:10.1029/2006JD008191.

    • Search Google Scholar
    • Export Citation
  • Betts, A. K., and P. Viterbo, 2005: Land surface, boundary layer, and cloud field coupling over the southwestern Amazon in ERA-40. J. Geophys. Res., 110, D14108, doi:10.1029/2004JD005702.

    • Search Google Scholar
    • Export Citation
  • Brimelow, J. C., J. M. Hanesiak, and R. Raddatz, 2010a: Validation of soil moisture simulations from the PAMII model, and an assessment of their sensitivity to uncertainties in soil hydraulic parameters. Agric. For. Meteor., 150, 100114.

    • Search Google Scholar
    • Export Citation
  • Brimelow, J. C., J. M. Hanesiak, R. Raddatz, and M. Hayashi, 2010b: Validation of ET estimates from the Canadian Prairie agrometeorological model for contrasting vegetation types and growing seasons. Can. Water Resour. J., 35, 209230.

    • Search Google Scholar
    • Export Citation
  • Brimelow, J. C., J. M. Hanesiak, and W. R. Burrows, 2011: On the surface-convection feedback during drought periods on the Canadian Prairies. Earth Interactions, 15. [Available online at http://EarthInteractions.org.]

    • Search Google Scholar
    • Export Citation
  • Burrows, W. R., and B. Kochtubajda, 2010: A decade of cloud-to-ground lightning in Canada: 1999-2008. Part 1: Flash density and occurrence. Atmos.–Ocean, 48, 177194.

    • Search Google Scholar
    • Export Citation
  • Chen, F., and Y. Zhang, 2009: On the coupling strength between the land surface and the atmosphere: From viewpoint of surface exchange coefficients. Geophys. Res. Lett., 36, L10404, doi:10.1029/2009GL037980.

    • Search Google Scholar
    • Export Citation
  • Crook, A., 1996: Sensitivity of moist convection forced by boundary layer processes to low-level thermodynamic fields. Mon. Wea. Rev., 124, 17671785.

    • Search Google Scholar
    • Export Citation
  • Cummins, K. L., R. B. Pyle, and G. Fournier, 1999: An integrated North American lightning detection network. Proc. 11th Int. Conf. on Atmospheric Electricity, Guntersville, AL, NASA, 218–221.

    • Search Google Scholar
    • Export Citation
  • Desai, A. R., K. J. Davis, C. J. Senff, S. Ismail, E. V. Browell, D. R. Stauffer, and B. P. Reen, 2006: A case study on the effects of heterogeneous soil moisture on mesoscale boundary-layer structure in the Southern Great Plains, U.S.A. Part I: Simple prognostic model. Bound.-Layer Meteor., 119, 195238.

    • Search Google Scholar
    • Export Citation
  • Dirmeyer, P. A., 1994: Vegetation stress as a feedback mechanism in midlatitude drought. J. Climate, 7, 14631483.

  • Dirmeyer, P. A., 2006: The hydrologic feedback pathway for land–climate coupling. J. Hydrometeor., 7, 857867.

  • Dirmeyer, P. A., C. A. Schloser, and K. L. Brubaker, 2009: Precipitation, recycling, and land memory: An integrated analysis. J. Hydrometeor., 10, 278288.

    • Search Google Scholar
    • Export Citation
  • Dominguez, F., and P. Kumar, 2008: Precipitation variability and ecoclimatological stability—A study using NARR data. Part I: Central U.S. Plains ecoregion. J. Climate, 21, 51655186.

    • Search Google Scholar
    • Export Citation
  • Doswell, C. A., 1987: The distinction between large-scale and mesoscale contribution to severe convection: A case study example. Wea. Forecasting, 2, 316.

    • Search Google Scholar
    • Export Citation
  • Findell, K. L., and E. A. B. Eltahir, 1999: Analysis of the pathways relating soil moisture and subsequent rainfall in Illinois. J. Geophys. Res., 104, 312 565312 574.

    • Search Google Scholar
    • Export Citation
  • Findell, K. L., and E. A. B. Eltahir, 2003a: Atmospheric controls on soil moisture–boundary layer interactions. Part I: Framework development. J. Hydrometeor., 4, 552569.

    • Search Google Scholar
    • Export Citation
  • Findell, K. L., and E. A. B. Eltahir, 2003b: Atmospheric controls on soil moisture–boundary layer interactions. Part II: Feedbacks within the continental United States. J. Hydrometeor., 4, 570583.

    • Search Google Scholar
    • Export Citation
  • Garrity, C. M., R. S. Cerveney, and E. A. Wentz, 2010: Vertical moisture profile characteristics of severe surface drought and surface wetness in the western United States: 1973–2002. Int. J. Climatol., 30, 894900.

    • Search Google Scholar
    • Export Citation
  • Gu, Y., J. F. Brown, J. P. Verdin, and B. Wardlow, 2007: A five-year analysis of MODIS NDVI and NDWI for grassland drought assessment over the central Great Plains of the United States. Geophys. Res. Lett., 34, L06407, doi:10.1029/2006GL029127.

    • Search Google Scholar
    • Export Citation
  • Hohenegger, C., P. Brockhaus, C. S. Bretherton, and C. Schär, 2009: The soil moisture–precipitation feedback in simulations with explicit and parameterized convection. J. Climate, 22, 50035020.

    • Search Google Scholar
    • Export Citation
  • Ji, L., and A. Peters, 2003: Assessing vegetation response to drought in the northern Great Plains using vegetation and drought indices. Remote Sens. Environ., 87, 8598.

    • Search Google Scholar
    • Export Citation
  • Jiang, X., G.-Y. Niu, and Z.-L. Yang, 2009: Impacts of vegetation and groundwater dynamics on warm season precipitation over the central United States. J. Geophys. Res., 114, D06109, doi:10.1029/2008JD010756.

    • Search Google Scholar
    • Export Citation
  • Juang, J.-Y., A. Porporato, P. C. Stoy, M. S. Siqueira, A. C. Oishi, M. Detto, H.-S. Kim, and G. G. Katul, 2007: Hydrologic and atmospheric controls on initiation of convective precipitation events. Water Resour. Res., 43, W03421, doi:10.1029/2006WR004954.

    • Search Google Scholar
    • Export Citation
  • Kim, Y., and G. Wang, 2007: Impact of vegetation feedback on the response of precipitation to antecedent soil moisture anomalies over North America. J. Hydrometeor., 8, 534550.

    • Search Google Scholar
    • Export Citation
  • Koster, R. D., and M. J. Suarez, 2004: Suggestions in the observational record of land–atmosphere feedback operating at seasonal time scales. J. Hydrometeor., 5, 567572.

    • Search Google Scholar
    • Export Citation
  • Koster, R. D., and Coauthors, 2004: Regions of strong coupling between soil moisture and precipitation. Science, 305, 11381140.

  • Koster, R. D., and Coauthors, 2006: GLACE: The Global Land–Atmosphere Coupling Experiment. Part I: Overview. J. Hydrometeor., 7, 590610.

    • Search Google Scholar
    • Export Citation
  • Latifovic, R., and Coauthors, 2005: Generating historical AVHRR 1 km baseline satellite data records over Canada suitable for climate change studies. Can. J. Remote Sens., 31, 324346.

    • Search Google Scholar
    • Export Citation
  • Lawrence, D. M., and J. M. Slingo, 2005: Weak land–atmosphere coupling strength in HadAM3: Moisture variability. J. Hydrometeor., 6, 670680.

    • Search Google Scholar
    • Export Citation
  • Liu, J., R. E. Stewart, and K. K. Szeto, 2004: Moisture transport and other hydrometeorological features associated with the severe 2000/01 drought over the western and central Canadian Prairies. J. Climate, 17, 305319.

    • Search Google Scholar
    • Export Citation
  • Liu, Z., M. Notaro, J. Kutzbach, and N. Liu, 2006: Assessing global vegetation–climate feedbacks from observations. J. Climate, 19, 787814.

    • Search Google Scholar
    • Export Citation
  • Matyas, C. J., and A. M. Carleton, 2010: Surface radar-derived convective rainfall associations with Midwest US land surface conditions in summer seasons 1999 and 2000. Theor. Appl. Climatol., 99, 315330.

    • Search Google Scholar
    • Export Citation
  • Mesinger, F., and Coauthors, 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87, 343360.

  • Mitchell, R. A. C., V. J. Mitchell, and D. W. Lowler, 2001: Response of wheat canopy CO2 and water gas-exchange to soil water content under ambient and elevated CO2. Global Change Biol., 7, 599611.

    • Search Google Scholar
    • Export Citation
  • Mkhabela, M., P. Bullock, M. Gervais, G. Finlay, and H. Sapirstein, 2010: Assessing indicators of agricultural drought impacts on spring wheat yield and quality on the Canadian prairies. Agric. For. Meteor., 150, 399410.

    • Search Google Scholar
    • Export Citation
  • Moller, A. R., 2001: Severe local storms forecasting. Severe Convective Storms, Meteor. Monogr., No. 28, Amer. Meteor. Soc., 433–480.

  • Myoung, B., and J. W. Nielsen-Gammon, 2010a: Sensitivity of monthly convective precipitation to environmental conditions. J. Climate, 23, 166188.

    • Search Google Scholar
    • Export Citation
  • Myoung, B., and J. W. Nielsen-Gammon, 2010b: The convective instability pathway to warm season drought in Texas. Part I: The role of convective inhibition and its modulation by soil moisture. J. Climate, 23, 44614473.

    • Search Google Scholar
    • Export Citation
  • Notaro, M., Z. Liu, and J. W. Williams, 2006: Observed vegetation climate feedbacks in the United States. J. Climate, 19, 763786.

  • Oglesby, R. J., S. Marshall, D. J. Erickson III, J. O. Roads, and F. R. Robertson, 2002: Thresholds in atmosphere–soil moisture interactions: Results from climate model studies. J. Geophys. Res., 107, 4224, doi:10.1029/2001JD001045.

    • Search Google Scholar
    • Export Citation
  • Pal, J. S., and E. A. B. Eltahir, 2001: Pathways relating soil moisture conditions to future summer rainfall within a model of the land–atmosphere system. J. Climate, 14, 12271242.

    • Search Google Scholar
    • Export Citation
  • Peppler, R. A., and P. J. Lamb, 1989: Tropospheric static stability and central North American growing season rainfall. Mon. Wea. Rev., 117, 11561180.

    • Search Google Scholar
    • Export Citation
  • Potvin, C. K., K. L. Elmore, and S. J. Weiss, 2010: Assessing the impacts of proximity sounding criteria on the climatology of significant tornado environments. Wea. Forecasting, 25, 921930.

    • Search Google Scholar
    • Export Citation
  • Raddatz, R. L., 1993: Prairie agroclimate boundary-layer model: A simulation of the atmosphere/crop-soil interface. Atmos.–Ocean, 4, 399419.

    • Search Google Scholar
    • Export Citation
  • Raddatz, R. L., and J. M. Hanesiak, 2008: Significant summer rainfall in the Canadian Prairie provinces: Modes and mechanisms 2000–2004. Int. J. Climatol., 28, 16071613.

    • Search Google Scholar
    • Export Citation
  • Schär, C., D. Luthi, and U. Beyerle, 1999: The soil-precipitation feedback: A process study with a regional climate model. J. Climate, 12, 722741.

    • Search Google Scholar
    • Export Citation
  • Schulze, E. D., 1986: Carbon dioxide and water vapor exchange in response to drought in the atmosphere and in the soil. Annu. Rev. Plant Physiol., 37, 247274.

    • Search Google Scholar
    • Export Citation
  • Schumann, M. R., and P. J. Roebber, 2010: The influence of upper-tropospheric potential vorticity on convective morphology. Mon. Wea. Rev., 138, 463474.

    • Search Google Scholar
    • Export Citation
  • Segal, M., R. W. Arritt, C. Clark, R. Rabin, and J. Brown, 1995: Scaling evaluation of the effect of surface characteristics on potential for deep convection over uniform terrain. Mon. Wea. Rev., 123, 383400.

    • Search Google Scholar
    • Export Citation
  • Seneviratne, S. I., T. Corti, E. L. Davin, M. Hirschi, E. B. Jaeger, I. Lehner, B. Orlowsky, and A. J. Teuling, 2010: Investigating soil moisture-climate interactions in a changing climate: A review. Earth Sci. Rev., 99, 125161.

    • Search Google Scholar
    • Export Citation
  • Siqueira, M. B., G. Katul, and A. Porporato, 2009: Soil moisture feedbacks on convection triggers: The role of soil–plant hydrodynamics. J. Hydrometeor., 10, 96112.

    • Search Google Scholar
    • Export Citation
  • Smith, S. B., and M. K. Yau, 1993a: The causes of severe convective outbreaks in Alberta. Part I: A comparison of a severe outbreak with two nonsevere events. Mon. Wea. Rev., 121, 10991125.

    • Search Google Scholar
    • Export Citation
  • Smith, S. B., and M. K. Yau, 1993b: The causes of severe convective outbreaks in Alberta. Part II: Conceptual model and statistical analysis. Mon. Wea. Rev., 121, 11261133.

    • Search Google Scholar
    • Export Citation
  • Stensrud, D. J., and J. M. Fritsch, 1993: Mesoscale convective systems in weakly forced large-scale environments. Part I: Observations. Mon. Wea. Rev., 121, 33263344.

    • Search Google Scholar
    • Export Citation
  • Strong, G. S., 1986: Synoptic to mesoscale dynamics of severe thunderstorm environments. A diagnostic study with forecasting implications. Ph.D thesis, University of Alberta, 345 pp.

    • Search Google Scholar
    • Export Citation
  • Taylor, C. M., P. P. Harris, and D. J. Parker, 2010: Impact of soil moisture on the development of a Sahelian mesoscale convective system: A case-study from the AMMA special observing period. Quart. J. Roy. Meteor. Soc., 136, 456470.

    • Search Google Scholar
    • Export Citation
  • Toumi, R., and X. Qie, 2004: Seasonal variation of lightning on the Tibetan Plateau: A Spring anomaly? Geophys. Res. Lett., 31, L04115, doi:10.1029/2003GL018930.

    • Search Google Scholar
    • Export Citation
  • Trier, S. B., F. Chen, and K. W. Manning, 2004: A study of convection initiation in a mesoscale model using high-resolution land surface initial conditions. Mon. Wea. Rev., 132, 29542976.

    • Search Google Scholar
    • Export Citation
  • Tucker, C. J., 1979: Red and photographic infrared linear combinations for monitoring vegetation. Remote Sens. Environ., 8, 127150.

  • Wang, S. Y., T. C. Chen, and J. Correia, 2009: Climatology of summer midtropospheric perturbations in the U.S. northern plains. Part I: Influence on northwest flow severe weather outbreaks. Climate Dyn., 36, 793810, doi:10.1007/s00382-009-0696-3.

    • Search Google Scholar
    • Export Citation
  • Wang, W., B. T. Anderson, N. Phillips, R. K. Kaufmann, C. Potter, and R. B. Myneni, 2006: Feedbacks of vegetation on summertime climate variability over the North American grasslands. Part I: Statistical analysis. Earth Interactions, 10. [Available online at http://EarthInteractions.org.]

    • Search Google Scholar
    • Export Citation
  • Yamada, H., 2008: Numerical simulations of the role of land surface conditions in the evolution and structure of summertime thunderstorms over a flat highland. Mon. Wea. Rev., 136, 173188.

    • Search Google Scholar
    • Export Citation
Save
  • Alfieri, J. G., P. Claps, P. D’Odorico, F. Laio, and T. M. Over, 2008: An analysis of the soil moisture feedback on convective and stratiform precipitation. J. Hydrometeor., 9, 280291.

    • Search Google Scholar
    • Export Citation
  • Anderson, M. C., J. M. Norman, J. R. Mecikalski, J. A. Otkin, and W. P. Kustas, 2007: A climatological study of evapotranspiration and moisture stress across the continental United States based on thermal remote sensing: 1. Model formulation. J. Geophys. Res., 112, D10117, doi:10.1029/2006JD007506.

    • Search Google Scholar
    • Export Citation
  • Beljaars, A. C. M., P. Viterbo, M. J. Miller, and A. K. Betts, 1996: The anomalous rainfall over the United States during July 1993: Sensitivity to land surface parameterization and soil moisture anomalies. Mon. Wea. Rev., 124, 362383.

    • Search Google Scholar
    • Export Citation
  • Betts, A. K., 2004: Understanding hydrometeorology using global models. Bull. Amer. Meteor. Soc., 85, 16731688.

  • Betts, A. K., 2007: Coupling of water vapor convergence, clouds, precipitation, and land surface processes. J. Geophys. Res., 112, D10108, doi:10.1029/2006JD008191.

    • Search Google Scholar
    • Export Citation
  • Betts, A. K., and P. Viterbo, 2005: Land surface, boundary layer, and cloud field coupling over the southwestern Amazon in ERA-40. J. Geophys. Res., 110, D14108, doi:10.1029/2004JD005702.

    • Search Google Scholar
    • Export Citation
  • Brimelow, J. C., J. M. Hanesiak, and R. Raddatz, 2010a: Validation of soil moisture simulations from the PAMII model, and an assessment of their sensitivity to uncertainties in soil hydraulic parameters. Agric. For. Meteor., 150, 100114.

    • Search Google Scholar
    • Export Citation
  • Brimelow, J. C., J. M. Hanesiak, R. Raddatz, and M. Hayashi, 2010b: Validation of ET estimates from the Canadian Prairie agrometeorological model for contrasting vegetation types and growing seasons. Can. Water Resour. J., 35, 209230.

    • Search Google Scholar
    • Export Citation
  • Brimelow, J. C., J. M. Hanesiak, and W. R. Burrows, 2011: On the surface-convection feedback during drought periods on the Canadian Prairies. Earth Interactions, 15. [Available online at http://EarthInteractions.org.]

    • Search Google Scholar
    • Export Citation
  • Burrows, W. R., and B. Kochtubajda, 2010: A decade of cloud-to-ground lightning in Canada: 1999-2008. Part 1: Flash density and occurrence. Atmos.–Ocean, 48, 177194.

    • Search Google Scholar
    • Export Citation
  • Chen, F., and Y. Zhang, 2009: On the coupling strength between the land surface and the atmosphere: From viewpoint of surface exchange coefficients. Geophys. Res. Lett., 36, L10404, doi:10.1029/2009GL037980.

    • Search Google Scholar
    • Export Citation
  • Crook, A., 1996: Sensitivity of moist convection forced by boundary layer processes to low-level thermodynamic fields. Mon. Wea. Rev., 124, 17671785.

    • Search Google Scholar
    • Export Citation
  • Cummins, K. L., R. B. Pyle, and G. Fournier, 1999: An integrated North American lightning detection network. Proc. 11th Int. Conf. on Atmospheric Electricity, Guntersville, AL, NASA, 218–221.

    • Search Google Scholar
    • Export Citation
  • Desai, A. R., K. J. Davis, C. J. Senff, S. Ismail, E. V. Browell, D. R. Stauffer, and B. P. Reen, 2006: A case study on the effects of heterogeneous soil moisture on mesoscale boundary-layer structure in the Southern Great Plains, U.S.A. Part I: Simple prognostic model. Bound.-Layer Meteor., 119, 195238.

    • Search Google Scholar
    • Export Citation
  • Dirmeyer, P. A., 1994: Vegetation stress as a feedback mechanism in midlatitude drought. J. Climate, 7, 14631483.

  • Dirmeyer, P. A., 2006: The hydrologic feedback pathway for land–climate coupling. J. Hydrometeor., 7, 857867.

  • Dirmeyer, P. A., C. A. Schloser, and K. L. Brubaker, 2009: Precipitation, recycling, and land memory: An integrated analysis. J. Hydrometeor., 10, 278288.

    • Search Google Scholar
    • Export Citation
  • Dominguez, F., and P. Kumar, 2008: Precipitation variability and ecoclimatological stability—A study using NARR data. Part I: Central U.S. Plains ecoregion. J. Climate, 21, 51655186.

    • Search Google Scholar
    • Export Citation
  • Doswell, C. A., 1987: The distinction between large-scale and mesoscale contribution to severe convection: A case study example. Wea. Forecasting, 2, 316.

    • Search Google Scholar
    • Export Citation
  • Findell, K. L., and E. A. B. Eltahir, 1999: Analysis of the pathways relating soil moisture and subsequent rainfall in Illinois. J. Geophys. Res., 104, 312 565312 574.

    • Search Google Scholar
    • Export Citation
  • Findell, K. L., and E. A. B. Eltahir, 2003a: Atmospheric controls on soil moisture–boundary layer interactions. Part I: Framework development. J. Hydrometeor., 4, 552569.

    • Search Google Scholar
    • Export Citation
  • Findell, K. L., and E. A. B. Eltahir, 2003b: Atmospheric controls on soil moisture–boundary layer interactions. Part II: Feedbacks within the continental United States. J. Hydrometeor., 4, 570583.

    • Search Google Scholar
    • Export Citation
  • Garrity, C. M., R. S. Cerveney, and E. A. Wentz, 2010: Vertical moisture profile characteristics of severe surface drought and surface wetness in the western United States: 1973–2002. Int. J. Climatol., 30, 894900.

    • Search Google Scholar
    • Export Citation
  • Gu, Y., J. F. Brown, J. P. Verdin, and B. Wardlow, 2007: A five-year analysis of MODIS NDVI and NDWI for grassland drought assessment over the central Great Plains of the United States. Geophys. Res. Lett., 34, L06407, doi:10.1029/2006GL029127.

    • Search Google Scholar
    • Export Citation
  • Hohenegger, C., P. Brockhaus, C. S. Bretherton, and C. Schär, 2009: The soil moisture–precipitation feedback in simulations with explicit and parameterized convection. J. Climate, 22, 50035020.

    • Search Google Scholar
    • Export Citation
  • Ji, L., and A. Peters, 2003: Assessing vegetation response to drought in the northern Great Plains using vegetation and drought indices. Remote Sens. Environ., 87, 8598.

    • Search Google Scholar
    • Export Citation
  • Jiang, X., G.-Y. Niu, and Z.-L. Yang, 2009: Impacts of vegetation and groundwater dynamics on warm season precipitation over the central United States. J. Geophys. Res., 114, D06109, doi:10.1029/2008JD010756.

    • Search Google Scholar
    • Export Citation
  • Juang, J.-Y., A. Porporato, P. C. Stoy, M. S. Siqueira, A. C. Oishi, M. Detto, H.-S. Kim, and G. G. Katul, 2007: Hydrologic and atmospheric controls on initiation of convective precipitation events. Water Resour. Res., 43, W03421, doi:10.1029/2006WR004954.

    • Search Google Scholar
    • Export Citation
  • Kim, Y., and G. Wang, 2007: Impact of vegetation feedback on the response of precipitation to antecedent soil moisture anomalies over North America. J. Hydrometeor., 8, 534550.

    • Search Google Scholar
    • Export Citation
  • Koster, R. D., and M. J. Suarez, 2004: Suggestions in the observational record of land–atmosphere feedback operating at seasonal time scales. J. Hydrometeor., 5, 567572.

    • Search Google Scholar
    • Export Citation
  • Koster, R. D., and Coauthors, 2004: Regions of strong coupling between soil moisture and precipitation. Science, 305, 11381140.

  • Koster, R. D., and Coauthors, 2006: GLACE: The Global Land–Atmosphere Coupling Experiment. Part I: Overview. J. Hydrometeor., 7, 590610.

    • Search Google Scholar
    • Export Citation
  • Latifovic, R., and Coauthors, 2005: Generating historical AVHRR 1 km baseline satellite data records over Canada suitable for climate change studies. Can. J. Remote Sens., 31, 324346.

    • Search Google Scholar
    • Export Citation
  • Lawrence, D. M., and J. M. Slingo, 2005: Weak land–atmosphere coupling strength in HadAM3: Moisture variability. J. Hydrometeor., 6, 670680.

    • Search Google Scholar
    • Export Citation
  • Liu, J., R. E. Stewart, and K. K. Szeto, 2004: Moisture transport and other hydrometeorological features associated with the severe 2000/01 drought over the western and central Canadian Prairies. J. Climate, 17, 305319.

    • Search Google Scholar
    • Export Citation
  • Liu, Z., M. Notaro, J. Kutzbach, and N. Liu, 2006: Assessing global vegetation–climate feedbacks from observations. J. Climate, 19, 787814.

    • Search Google Scholar
    • Export Citation
  • Matyas, C. J., and A. M. Carleton, 2010: Surface radar-derived convective rainfall associations with Midwest US land surface conditions in summer seasons 1999 and 2000. Theor. Appl. Climatol., 99, 315330.

    • Search Google Scholar
    • Export Citation
  • Mesinger, F., and Coauthors, 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87, 343360.

  • Mitchell, R. A. C., V. J. Mitchell, and D. W. Lowler, 2001: Response of wheat canopy CO2 and water gas-exchange to soil water content under ambient and elevated CO2. Global Change Biol., 7, 599611.

    • Search Google Scholar
    • Export Citation
  • Mkhabela, M., P. Bullock, M. Gervais, G. Finlay, and H. Sapirstein, 2010: Assessing indicators of agricultural drought impacts on spring wheat yield and quality on the Canadian prairies. Agric. For. Meteor., 150, 399410.

    • Search Google Scholar
    • Export Citation
  • Moller, A. R., 2001: Severe local storms forecasting. Severe Convective Storms, Meteor. Monogr., No. 28, Amer. Meteor. Soc., 433–480.

  • Myoung, B., and J. W. Nielsen-Gammon, 2010a: Sensitivity of monthly convective precipitation to environmental conditions. J. Climate, 23, 166188.

    • Search Google Scholar
    • Export Citation
  • Myoung, B., and J. W. Nielsen-Gammon, 2010b: The convective instability pathway to warm season drought in Texas. Part I: The role of convective inhibition and its modulation by soil moisture. J. Climate, 23, 44614473.

    • Search Google Scholar
    • Export Citation
  • Notaro, M., Z. Liu, and J. W. Williams, 2006: Observed vegetation climate feedbacks in the United States. J. Climate, 19, 763786.

  • Oglesby, R. J., S. Marshall, D. J. Erickson III, J. O. Roads, and F. R. Robertson, 2002: Thresholds in atmosphere–soil moisture interactions: Results from climate model studies. J. Geophys. Res., 107, 4224, doi:10.1029/2001JD001045.

    • Search Google Scholar
    • Export Citation
  • Pal, J. S., and E. A. B. Eltahir, 2001: Pathways relating soil moisture conditions to future summer rainfall within a model of the land–atmosphere system. J. Climate, 14, 12271242.

    • Search Google Scholar
    • Export Citation
  • Peppler, R. A., and P. J. Lamb, 1989: Tropospheric static stability and central North American growing season rainfall. Mon. Wea. Rev., 117, 11561180.

    • Search Google Scholar
    • Export Citation
  • Potvin, C. K., K. L. Elmore, and S. J. Weiss, 2010: Assessing the impacts of proximity sounding criteria on the climatology of significant tornado environments. Wea. Forecasting, 25, 921930.

    • Search Google Scholar
    • Export Citation
  • Raddatz, R. L., 1993: Prairie agroclimate boundary-layer model: A simulation of the atmosphere/crop-soil interface. Atmos.–Ocean, 4, 399419.

    • Search Google Scholar
    • Export Citation
  • Raddatz, R. L., and J. M. Hanesiak, 2008: Significant summer rainfall in the Canadian Prairie provinces: Modes and mechanisms 2000–2004. Int. J. Climatol., 28, 16071613.

    • Search Google Scholar
    • Export Citation
  • Schär, C., D. Luthi, and U. Beyerle, 1999: The soil-precipitation feedback: A process study with a regional climate model. J. Climate, 12, 722741.

    • Search Google Scholar
    • Export Citation
  • Schulze, E. D., 1986: Carbon dioxide and water vapor exchange in response to drought in the atmosphere and in the soil. Annu. Rev. Plant Physiol., 37, 247274.

    • Search Google Scholar
    • Export Citation
  • Schumann, M. R., and P. J. Roebber, 2010: The influence of upper-tropospheric potential vorticity on convective morphology. Mon. Wea. Rev., 138, 463474.

    • Search Google Scholar
    • Export Citation
  • Segal, M., R. W. Arritt, C. Clark, R. Rabin, and J. Brown, 1995: Scaling evaluation of the effect of surface characteristics on potential for deep convection over uniform terrain. Mon. Wea. Rev., 123, 383400.

    • Search Google Scholar
    • Export Citation
  • Seneviratne, S. I., T.