1. Introduction
Root-zone soil moisture in vegetated regions has a significant influence on evapotranspiration rates (McPherson 2007; Alfieri et al. 2008) and latent and sensible heat exchange (Dirmeyer et al. 2000; Basara and Crawford 2002). Inclusion of soil moisture in seasonal climate predictions has been shown to increase seasonal precipitation predictive accuracy in some cases (McPherson et al. 2004; Meng and Quiring 2010a; Koster et al. 2011). Meng and Quiring (2010a) found that spring soil moisture influenced summer precipitation in the Community Atmosphere Model, version 3.0 (CAM3.0). Soil moisture is thought to influence precipitation and temperature on seasonal scales through soil moisture memory (Wu and Dickinson 2004), such that anomalously low soil moisture reduces the amount of moisture available for precipitation recycling. Under these conditions, locally sourced precipitation within 200 km of initiation is less likely than when soil moisture is normal or wetter than normal (Dirmeyer et al. 2009). Despite the overwhelming evidence of soil moisture–precipitation coupling in modeling studies, few investigations have been able to conclusively show soil moisture impact on subsequent precipitation using observations of both soil moisture and precipitation. This is partly due to the lack of spatially and temporally extensive in situ soil moisture observations (Robock et al. 2000; Seneviratne et al. 2010), as well as the strong autocorrelation of precipitation, which can mask any signal of soil moisture forcing (Wei et al. 2008).
More recently, the focus has shifted from seasonal relationships to investigations attempting to connect soil moisture to subsequent precipitation on much shorter time scales. Specifically, soil moisture feedback to convective initiation or the likelihood of convective precipitation has been examined. Ek and Holtslag (2004) explored the feedback of soil moisture on planetary boundary layer (PBL) convective cloud development and found that relative humidity at the PBL is related to surface evaporation and partitioning of energy between sensible and latent heat flux. Taylor et al. (2007) found that strong horizontal gradients in observed sensible heat flux, forced by soil moisture gradients, can generate sea breeze–like circulations in the lower atmosphere. These mesoscale circulations can provide the focus for moist convection, which generates precipitation preferentially over dry soils adjacent to wet soils. Garcia-Carreras et al. (2011) showed that initiation of mesoscale convection is found preferentially on the warm (dry) side of dry–wet soil boundaries. Myoung and Nielsen-Gammon (2010) found that anomalously wet soils were negatively correlated with convective inhibition in the southern Great Plains, which suggests that wet soils are associated with an increased likelihood of convective initiation in this region.
Several factors obscure the coupling relationship between soil moisture and convective precipitation. One issue is the different mechanisms through which soil moisture can influence convection. Taylor et al. (2011) found that convective precipitation fell preferentially over dry soils in the Sahel region of Africa. Their explanation was that anomalously dry soils increase sensible heat flux and surface heating, which destabilizes the atmospheric profile in the PBL and creates an area of local convergence, increasing the likelihood of convection and precipitation over the relatively dry soils. In contrast, Kang and Bryan (2011) found that greater cooling and moistening in the PBL over relatively moist surfaces is associated with earlier convective initiation than over relatively dry surfaces. Brimelow et al. (2011) compared convection-related indices with observations of normalized difference vegetation index (NDVI), a proxy for evapotranspiration. They found that reduced latent heat flux resulting from (inferred) low soil moisture and stressed vegetation were largely responsible for a deeper, drier boundary layer. In this case, anomalously dry soils were connected to decreased convective available potential energy (CAPE), increased lifting condensation level, and increased convective inhibition. Another potential confounding factor that influences the sign and strength of land–atmosphere feedbacks is the synoptic–dynamic environment of a particular region. For example, Frye and Mote (2010) show that the moisture and energy flux from the Gulf of Mexico into the southern Great Plains have a strong impact on the overall convective environment and the strength of convection.
Another issue with quantifying soil moisture–precipitation coupling is the indirect connection between the two variables. Soil moisture does not directly influence precipitation, but instead modulates latent and sensible heat flux in the near-surface atmosphere (Pielke 2001; Gu et al. 2006; Wei and Dirmeyer 2012). The corresponding energy and moisture fluxes in the near-surface atmosphere then have the potential to impact convective initiation and precipitation (Taylor et al. 2011; Jones and Brunsell 2009; Matyas and Carleton 2010; Santanello et al. 2011; Santanello and Peters-Lidard 2013). Despite several studies examining land surface impacts on precipitation, soil moisture proxies such as land surface temperature from satellites or reanalysis data are most frequently used. We employ in situ soil moisture estimates to explore soil moisture–precipitation coupling in Oklahoma. Our in situ observations provide an improvement over satellite remote sensing–based soil moisture products, which are affected by signal attenuation by vegetation (Crow et al. 2001; Wagner et al. 2007), leading to significant spatial variability in product accuracy (Jackson et al. 2010; Ford et al. 2014). The primary disadvantage of in situ data is their limited spatial representativeness. This study examines whether convective precipitation falls preferentially over wet or dry soils and documents the atmospheric and land surface conditions prior to the occurrence of convection. Soil moisture anomalies are investigated in the context of modification of subsequent air temperature and humidity anomalies as well as CAPE during the warm season (May–September) from 2003 to 2012.
2. Data and methods
a. Study area
Oklahoma is located in the southern Great Plains and experiences a significant west–east precipitation gradient and north–south temperature gradient (Figs. 1a,b). Land cover ranges from grassland and forest in the east to cropland and scrubland in the central and western portions of the state. Oklahoma and the larger southern Great Plains region have been the focal point of several land–atmosphere interaction studies that have suggested that anomalously dry or wet soils can influence precipitation on hourly (Frye and Mote 2010), monthly, and seasonal time scales (McPherson et al. 2004; Meng and Quiring 2010b). The southern Great Plains, especially Oklahoma, also contains the densest network of continuously monitoring in situ matric potential stations in the United States. This study uses in situ soil moisture data, estimated from matric potential from 2003 to 2012, to investigate its relationship with warm season convective precipitation.
Spatial distributions of annual average (a) precipitation (mm) and (b) temperature (°C). The black dots represent Oklahoma Mesonet stations with soil moisture data used in this study. Data are taken from the PRISM (http://prism.oregonstate.edu) precipitation and temperature datasets.
Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0005.1
b. Soil moisture data
The Oklahoma Mesonet (www.mesonet.org; Illston et al. 2008) is a statewide monitoring network of more than 100 stations. This study utilizes in situ volumetric water content (cm3 cm−3) from 113 Oklahoma Mesonet stations (Fig. 1). Volumetric water content of the soil is estimated using the thermal matric potential that is measured by Campbell 229-L heat dissipation sensors at 5, 25, and 60 cm. Physical soil moisture properties were recently updated at each Oklahoma Mesonet site, and soil moisture observations from the Campbell 229-L sensors were evaluated with gravimetric samples (Scott et al. 2013). Root-mean-square difference between direct measurements of volumetric water content and those reported by the 229-L sensors varied from 0.08 to 0.05 cm3 cm−3, with an overall network average of 0.05 cm3 cm−3 (Scott et al. 2013). The Oklahoma Mesonet is the first mesoscale soil moisture observation network to quantify the network-wide uncertainty in their observations. The errors in the Oklahoma Mesonet observations are generally less than those of satellite- and model-derived soil moisture (Xia et al. 2014). For example, Jackson et al. (2010) showed that the root-mean-square difference between in situ soil moisture observations and the AMSR-E satellite remote sensing soil moisture product varied from 0.02 to 0.22 cm3 cm−3, depending strongly on the validation location and retrieval algorithm. Over Oklahoma, Jackson et al. (2007) compared AMSR-E to in situ data from 16 stations and found root-mean-square difference values ranged between 0.04 and 0.10 cm3 cm−3, again depending on the retrieval algorithm. Ford et al. (2014) showed that standardized soil moisture mean absolute differences between Oklahoma Mesonet observations and remote sensing retrievals from the Soil Moisture and Ocean Salinity (SMOS) platform varied from 0.14 to 0.43 cm3 cm−3. Satellite soil moisture products can provide an accurate depiction of large-scale soil moisture variability; however, extensive validation is necessary to estimate the accuracy of the satellite-derived soil moisture products.
This study uses soil moisture observations from the Oklahoma Mesonet at 5-cm depth (Illston et al. 2008) that were measured at 0000 central standard time (CST). Observations later in the morning (e.g., 0900 CST) and at a higher temporal resolution would have been preferred for this study; however, these data were not available. An empirical cumulative distribution function was determined using all (0000 CST) volumetric water content measurements for each station between 2003 and 2012. The empirical cumulative distribution was generated for all daily volumetric water content measurements in a given month and was subsequently used to convert the soil moisture observations to percentiles. We remove the soil moisture seasonal cycle by using a separate distribution function for each calendar month (Illston et al. 2008). Percentiles provide a standardized measurement of relative soil wetness with respect to the soil moisture conditions experienced at each station during the study period. In this way, the soil moisture value in May with a percentile of 100 is the wettest soil moisture observation of any May over the entire study period at that particular station. These station-based soil moisture percentiles were gridded at a 0.25° resolution using the average of all stations within each grid cell. Figure 2 shows the spatial distribution of Mesonet sites with respect to the 0.25° grid cells. Over the entire study region (431 grid cells), 113 cells have at least one Mesonet station. Out of these, six grid cells have two stations and three have three stations.
Soil moisture grid cells used in the study. Grid cells are color coded based on the number of Oklahoma Mesonet stations in each cell.
Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0005.1
c. Precipitation event identification
To capture the spatial variability and relatively short time scales of convective precipitation, we apply the methods of Taylor et al. (2012). They used the Climate Prediction Center (CPC) morphing method (CMORPH; Joyce et al. 2004) dataset to identify daily afternoon precipitation events in their global study of the linkage between local soil moisture conditions and the likelihood of precipitation. CMORPH merges satellite passive microwave rainfall estimates and thermal infrared data to provide global precipitation estimates every 3 h from 2002 to 2014. CMORPH takes advantage of the high spatial and temporal sampling of the geostationary data to derive motion vectors that are then used to propagate the less frequent passive microwave rainfall retrievals to produce the 3-hourly precipitation estimates.
Similar to Taylor et al. (2012), the CMORPH 3-hourly precipitation estimates are totaled over each 0.25° grid box in Oklahoma for each day from 1800 to 0300 UTC (1200–2100 CST) in the afternoon and from 0600 to 1800 UTC (0000–1200 CST) on the event-day morning. Afternoon precipitation maxima are identified as grid boxes with afternoon accumulations greater than 3 mm. To avoid biases due to morning precipitation before a convective event and after the soil moisture observation, events are excluded when precipitation accumulations are greater than 1 mm between midnight and noon (i.e., 0000–1200 CST) the morning of an event. The daily maxima are then sorted by total afternoon accumulations and the grid cell with the maximum accumulation is identified as the convective event. A 1.25° × 1.25° box is centered at the location of maximum precipitation for each event and the location of the maximum and minimum precipitation grid cells within this box are identified. These locations are then related to the corresponding soil moisture value from the gridded soil moisture observations. This analysis was originally intended by Taylor et al. (2012) to determine whether afternoon precipitation occurs more frequently when local soil moisture at the event location is higher or lower than the surrounding area; however, it also allows for analysis of whether precipitation events occur preferentially over wetter or dryer soils for a given soil moisture station location. Because events are defined within a 1.25° × 1.25° box, any precipitation maxima whose surrounding box overlaps another event box with greater accumulation is excluded to maximize the independence of precipitation events.
d. Synoptic environment
Land–atmosphere interactions occur when the local land surface and subsurface conditions influence the moisture and energy budgets of the overlying atmosphere (Pielke 2001). Therefore, soil moisture impacts on convective precipitation could be more noticeable depending on the overall atmospheric conditions (i.e., synoptic forcing is weak). We account for synoptic conditions by classifying each day into one of four types. The classification system we employed is adopted from Frye and Mote (2010) and uses the environmental lapse rate and presence or absence of the Great Plains low-level jet (LLJ) to characterize synoptic conditions. Both the lapse rate and presence or absence of the LLJ were determined using daily 0900 CST data from the North American Regional Reanalysis (NARR; Mesinger et al. 2006) dataset. The NARR dataset is a high-resolution, high-frequency atmospheric and land surface hydrology dataset, which covers from 1979 to the present and provides several meteorological and hydrological variables eight times daily at a 32-km resolution over the northern Lambert conformal conic grid (North America). The lapse rate was calculated from the atmospheric temperature profile of the NARR grid cell overlying each event between 850 and 700 hPa. If the lapse rate is less than 6.0°C km−1, the synoptic environment is considered stable or synoptically benign (SB; Frye and Mote 2010). If the lapse rate is greater than 6.0°C km−1, the synoptic environment is considered unstable and is classified as synoptically prime (SP).
The LLJ is a large-scale feature that influences weather patterns throughout the central United States (Bonner 1968) and is the primary mechanism by which moisture is advected from the Gulf of Mexico into the south and central Great Plains (Higgins et al. 1997; Wu and Raman 1998). The LLJ can impact the stability of the lower atmosphere and the convective environment, potentially masking the influence of local moisture conditions. Here, the LLJ is classified as either present or absent using daily 0900 CST, 850-hPa level wind vector data from the NARR dataset. Similar to Frye and Mote (2010), we determine the LLJ is present if vector winds with speeds in excess of 12 m s−1 (Bonner 1968) originating from the Gulf of Mexico influence the majority of Oklahoma. Majority in this case means more than 75% of grid cells have a southerly or southeasterly vector wind in excess of 12 m s−1. The synoptic environment classification system results in four classes: synoptically prime–LLJ present (SP-LLJ), synoptically benign–LLJ present (SB-LLJ), synoptically prime–LLJ absent (SP-noLLJ), and synoptically benign–LLJ absent (SB-noLLJ). Every day with a precipitation event is classified into one of these four classes. This procedure allows us to separate events by the overall synoptic and dynamic conditions and focus on events without the impact of synoptic-scale atmospheric processes that could confound the land surface relationship.
e. Convection indices
The convective environment associated with each precipitation event is also characterized using surface CAPE calculated from (1800 UTC) NARR data. High values of CAPE represent unstable atmospheric conditions with high potential for moist convection. Taylor and Lebel (1998) suggest that soil moisture anomalies can have a significant influence on CAPE. The daily CAPE values are converted to anomalies (J kg−1) using the mean of all days (event and nonevent) during the event month between 2003 and 2012. For example, the CAPE value associated with an event on 1 June 2005 is converted to a departure from normal by subtracting the mean CAPE of all June days between 2003 and 2012.
f. Temperature and humidity
Along with convective environments, we also define near-surface atmospheric temperature and specific humidity from the near surface (1000–900 hPa) before each event. Temperature and specific humidity are from 1800 UTC NARR data and they are compared to soil moisture and CAPE. Because both temperature and humidity exhibit considerable spatial variability in Oklahoma, both variables were converted to z scores by subtracting the mean of all days during the event month and dividing by the standard deviation of all days during the event month between 2003 and 2012.
3. Results
a. Synoptic classes
A total of 1697 events are classified between May and September 2003–12. Of the 1697 events, 353 occurred over grid cells with available soil moisture observations. Therefore, we only analyze these 353 events; Fig. 3 shows the spatial distribution of the precipitation events across Oklahoma. No significant spatial patterns of event occurrence are present, and therefore, the spatial gradients of temperature, precipitation, and vegetation are assumed to not impact the results. The number of events in each class is provided in Table 1. Nearly 70% of convective events identified occur on days when the LLJ is present, consistent with the results of Frye and Mote (2010). The monthly distribution of events of each type (Fig. 4) shows that events occurring in a synoptically prime environment tend to occur more frequently in July and August, while those occurring in a synoptically benign environment show a relative minimum in July.
Map of precipitation events occurring in each grid cell. Grid cells are color coded based on the number events in each cell.
Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0005.1
Number of afternoon precipitation events in each synoptic environment class and the percentage of overall total precipitation events in each class.
Mean monthly distribution of afternoon precipitation events during the study period (2003–12). Event frequencies are delineated by the synoptic class.
Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0005.1
b. Soil moisture magnitude
The location (grid cell) with the maximum precipitation was identified for each precipitation event and is used to identify the soil moisture conditions in that location. For each synoptic class (SP-LLJ, SB-LLJ, SP-noLLJ, and SB-noLLJ), the soil moisture conditions associated with each precipitation event are examined. The distributions of 5-cm soil moisture percentiles for each class are shown in Figs. 5a–d. Soil moisture during the morning of precipitation events in SP-LLJ and SB-LLJ classes (Figs. 5a,b) are more frequently drier than normal (<50%), while events in SP-noLLJ and SB-noLLJ classes seem to occur more frequently over wetter-than-normal soils (>50%).
Distributions of 5-cm soil moisture percentiles for all (a) SP-LLJ, (b) SB-LLJ, (c) SP-noLLJ, and (d) SB-noLLJ events.
Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0005.1
The statistical significance of soil moisture conditions underlying events in each class is evaluated against soil moisture conditions simply occurring by chance. The evaluation compares the number of events occurring over wet (dry) soils in each synoptic–dynamic class to the number of days with wet (dry) soils in a randomly selected sample of event and nonevent days. For clarification purposes, afternoon precipitation could occur on nonevent days; however, the precipitation would have to occur before noon or occur on an afternoon that followed a morning with precipitation. We employed a bootstrap resampling procedure with 10 000 iterations to randomly select a sample of n (event and nonevent) days. The number of days in each of the 10 000 samples was equal to the number of events in each synoptic class (e.g., 88 for SP-LLJ), and synoptic conditions of days in each sample matched conditions of the respective event class. That is, all 10 000 randomly selected days in the SB-LLJ sample were also SB-LLJ days between May and September 2003–12. Soil moisture observations for each of the days in the sample were taken from a randomly chosen grid cell. For each iteration, we tallied the frequency of days with relatively wet and relatively dry soil moisture conditions. The respective frequencies of relatively dry (<0.5) and relatively wet (>0.5) soil moisture events in each synoptic class were compared to the distribution of frequencies generated from the bootstrapping procedure, and the likelihood p of the event frequencies occurring by chance was calculated. For this evaluation, if the number of events occurring over dry soils in the SB-LLJ class represented the 96th percentile of the SB-LLJ bootstrapped sample, then the probability (p value) of that number of dry soil events occurring by chance is 0.04 or 4%.
Table 2 shows the probability of the frequency of events occurring over wet and dry soils for each synoptic class. Probabilities are calculated separately for event frequency over wet and dry soils. As Table 2 shows, the preference for SP-LLJ and SB-LLJ events to occur over dry soils is statistically significant at the 95% level, as is the preference for SP-noLLJ and SB-noLLJ events to occur over wet soils. Afternoon precipitation falls preferentially over both dry and wet soils, depending on the synoptic and dynamic conditions. Interestingly, the divide between precipitation falling over drier (wetter)-than-normal soils coincides with the presence (absence) of the LLJ. The impact of the LLJ on preference for convection over wetter- or drier-than-normal soils is discussed further in section 4.
Probability of events over wet and dry soils occurring by chance. The p values are generated by comparing frequency of events occurring over wet/dry soils to distributions of wet and dry day frequencies from a resampling bootstrapping procedure. The p values are reported by the synoptic class.
Statistically significant preferences for afternoon precipitation over both wet and dry soils are mechanistically consistent with potential negative and positive soil moisture feedbacks; however, our results do not provide causal links between soil moisture and precipitation. Several confounding factors can impact the relationship observed here, including the presence of larger-scale (e.g., frontal systems) atmospheric variability and day-to-day precipitation autocorrelation. Taylor et al. (2012) accounted for large-scale atmospheric persistence when determining if precipitation fell preferentially over dry or wet soils by comparing the soil moisture anomaly at the point of maximum precipitation to that of surrounding, precipitation minima grid cell(s). In this way, they were able to conclude that precipitation fell preferentially over soil patches that were drier than adjacent patches. We also attempt to account for larger-scale atmospheric persistence by comparing soil moisture at precipitation event maxima to adjacent, minima grid cells. Figures 6a–d show histograms of differences in percentiles between maximum and minimum grid cells for each event class. The histograms do not show a strong preference for soil moisture at precipitation maxima to be consistently drier or wetter than the adjacent grid cells. We tested the statistical significance of the frequency of differences in which the soil moisture maxima are wetter than the minima and those in which the minima are wetter than the maxima using the same bootstrapping methodology. For all event classes, the maxima − minima differences for both wetter and drier conditions were not statistically significantly different from that which occurs by chance. The negligible differences between soil moisture at precipitation maxima and adjacent minima for the events suggest that spatial autocorrelation of soil moisture percentiles in Oklahoma is quite high. Indeed, Ford and Quiring (2014) have shown that soil moisture spatial autocorrelation coefficients in Oklahoma exceed 0.5 at a distance of 200 km. Therefore, it is not surprising that the soil moisture percentiles associated with local precipitation minima were nearly identical to those associated with local precipitation maxima. This result also indicates that, unlike Taylor et al. (2012), we cannot discount the impact of larger-scale atmospheric persistence on our results.
Distributions of absolute differences between soil moisture percentiles underlying max and min rain rates. Distributions are separated by synoptic class.
Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0005.1
c. Temperature and humidity anomalies
Soil moisture does not directly influence precipitation, but instead modifies the partitioning of latent and sensible heat, impacting both temperature (Fischer et al. 2007; Teuling and Seneviratne 2008) and humidity (Basara and Crawford 2002; Roundy et al. 2013). Specifically, temperature is lower (higher) and humidity is higher (lower) over relatively wet (dry) soils. Increased evaporation from wet soils supplies additional moisture to the atmosphere for convection, providing a physical mechanism for positive soil moisture–precipitation feedback (Koster et al. 2003; Findell et al. 2011). Conversely, increased sensible heat from dry soils supplies uplift and instability for convection, providing a physical mechanism for negative soil moisture–precipitation feedback (Taylor et al. 2011). Our results in section 3b show afternoon precipitation falls preferentially over dry and wet soils, based on the overall synoptic and dynamic conditions. Here, we examine the relationship between soil moisture and near-surface atmospheric humidity and temperature to determine the potential of soil moisture feedback.
We use 1800 UTC (1200 CST) NARR near-surface (1000–900 hPa) temperature T and specific humidity q z scores to examine the relationship between soil moisture and the near-surface atmosphere. We first group event soil moisture percentiles into 10 bins along the soil moisture percentile gradient from 0 to 100. We then compare the average q and T anomaly of the soil moisture bins to the frequency of events in each bin. Figures 7a–d show the frequency of events in binned soil moisture percentiles and the mean near-surface specific humidity anomaly corresponding to events within each bin. We show the individual fluctuations of near-surface q and T corresponding with bins of soil moisture percentiles; however, the sample sizes in each bin are not large enough for the fluctuations to be significant. Therefore, we only remark on the general q and T trend as percentiles range from dry to wet. The general T z-score trends are decreasing as soil moisture increases for all event classes. Correlation coefficients r between soil moisture and temperature range from −0.75 (SP-noLLJ) to −0.88 (SB-LLJ) (α < 0.05) and are statistically significant for all four classes. There is a positive trend in q z scores as soil moisture percentiles increase; however, the slope of the trend varies more between classes than the T trends. Negative correlations between soil moisture and temperature are contrasted by positive soil moisture–humidity correlations. Correlations between soil moisture and humidity are statistically significant for all classes except SP-noLLJ, and coefficients (r) are decreased from soil moisture–temperature correlations, ranging from 0.20 (SP-noLLJ) to 0.73 (SP-LLJ).
Frequency of (a) SP-LLJ, (b) SB-LLJ, (c) SP-noLLJ, and (d) SB-noLLJ events grouped into soil moisture percentiles bins of 10. Near-surface humidity (blue line) and temperature (red line) z scores are averaged for each percentile bin and plotted.
Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0005.1
Figure 7 shows that soil moisture anomalies correlate well with near-surface atmospheric humidity and temperature anomalies. SP-LLJ and SB-LLJ events occur most frequently over relatively dry soils (Fig. 5, Table 2), which correspond with below-normal near-surface humidity for SP-LLJ events and above-normal near-surface temperatures for SP-LLJ and SB-LLJ events. This corroborates the findings of Taylor and Ellis (2006) and Taylor et al. (2011), which suggested that dry soils increase energy partitioned to sensible heat, decreasing low-level atmospheric stability and increasing uplift. In contrast, the majority of SB-noLLJ events occur over wetter-than-normal soils, which corresponds with increased near-surface atmospheric humidity and decreased temperature. This in turn is consistent with a positive land–atmosphere feedback such that the probability of precipitation is increased through invigorated latent heat flux and decreased lifting condensation level height (Pal and Eltahir 2001; Santanello et al. 2011). The connection between soil moisture percentiles and subsequent near-surface temperature and humidity z scores suggests a physical link through which soil moisture could potentially impact atmospheric conditions. The results from these statistical analyses are consistent with negative and positive feedback mechanisms proposed in previous studies. However, the aforementioned confounding factors make it impossible to provide a causal link between soil moisture anomalies and precipitation based solely on these results.
d. Convective environment
Temperature and humidity anomalies can decrease atmospheric stability, increasing the likelihood of moist convection in the planetary boundary layer. The results in the previous section suggest interactions between soil moisture and both the overlying atmospheric humidity and temperature, providing evidence of potential land–atmosphere coupling. Here we examine the covariation of soil moisture, near-surface humidity and temperature, and CAPE for each event. We use gridded CAPE, which represents general atmospheric instability and potential for moist convection, from 1800 UTC NARR data to characterize the atmosphere prior to each event. We examine the relationship between soil moisture underlying each event and the corresponding 1800 UTC q and T z scores.
Figures 8a–d show scatter points separated by dry and wet soils, plotted in dual (1000–900 mb) T–q z-score space. The size of the symbol is proportional to the 1800 UTC CAPE anomaly corresponding with each event. Events over dry soils are represented by red circles, and events over wet soils are represented by blue squares. This provides a means to visualize how soil moisture and CAPE covary in T–q space. The most noticeable pattern is the larger CAPE anomalies preceding events with higher-than-normal near-surface humidity. This is intuitive as increased humidity in the lower troposphere, whether from local evapotranspiration or advection, directly increases CAPE. This pattern is evident for SP-noLLJ events as well; however, this is the only class without a significant relationship between soil moisture and specific humidity (Fig. 7c).
Scatterplots of humidity and temperature anomalies for (a) SP-LLJ, (b) SB-LLJ, (c) SP-noLLJ, and (d) SB-noLLJ events. Each point represents one event that is separated into dry soil events (red circles) and wet soil events (blue squares) and is scaled based on the corresponding CAPE anomaly (J kg−1). The gray dashed lines are used to separate q and T into quadrants.
Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0005.1
Given the strong relationship between soil moisture and near-surface temperature anomalies (Figs. 7a–d), one would expect consistently warmer (cooler)-than-normal temperatures over drier (wetter)-than-normal soils. Figures 8a and 8b show the majority of SP-LLJ and SB-LLJ events over both dry and wet soils occur more frequently in conditions with warmer-than-normal near-surface temperatures. This pattern is somewhat unexpected for events over wet soils, as high soil moisture generally coincides with cooler-than-normal temperatures. One possible explanation for warm air over wetter-than-normal soils is the presence of the LLJ. During the warm season, the Great Plains LLJ is mostly southerly to southwesterly (Whiteman et al. 1997). In the Great Plains, this corresponds with warm advection from the southwest, increasing low-level air temperature. Despite wetter-than-normal soils, increased precipitation events and increased CAPE could be attributed to the presence of the southerly-to-southwesterly LLJ prior to and during SP-LLJ and SB-LLJ events.
While there are clearly variations in CAPE anomalies that appear to be at least partially linked to the dynamical environment, it is also noticeable that in LLJ conditions, dry soils correspond with larger CAPE anomalies than wet soils. For a given positive q anomaly in LLJ events, T anomalies and thus CAPE anomalies are on average larger when the soils are dry relative to when they are wet. Conversely, both event classes with no LLJ have larger CAPE anomalies for wet soils relative to dry soils, which occur even for weak positive or slightly negative T anomalies. This suggests the importance of the increased q (T) anomalies in the larger CAPE anomalies over wet (dry) soils.
4. Discussion
Soil moisture–precipitation interactions have been a major avenue of hydroclimatic research for decades. Previous studies have found evidence of a wet-positive soil moisture feedback in which anomalously wet soils lead to elevated latent heat flux, increased locally sourced moisture into the low-level atmosphere, destabilization of the lower atmosphere, and a higher likelihood of convective initiation (Pielke 2001; Findell and Eltahir 2003; Pal and Eltahir 2003; Koster et al. 2004; Ferguson and Wood 2011). In contrast, others have found that anomalously dry soils can impact convective initiation more strongly than wet soils through increased sensible heat flux and a corresponding decreased lower atmospheric stability and faster PBL growth (Taylor et al. 2007; Santanello et al. 2011; Taylor et al. 2012).
Our results do not show unequivocal evidence for strong soil moisture–precipitation coupling in Oklahoma, but instead suggest that afternoon precipitation is more likely to fall over drier- or wetter-than-normal soils depending on the synoptic and dynamic conditions. More specifically, events that occur with the LLJ present tend to occur more frequently over drier-than-normal soils, while those without an LLJ occur over wetter-than-normal soils. These preferences are statistically significant (Table 2); however, our results merely suggest possible soil moisture feedback to precipitation and do not provide a causal link between the two. The common denominator between events occurring over drier (wetter) soils is the presence (absence) of the LLJ, which is consistent with the findings of Frye and Mote (2010). This could partly explain the preference for SP-LLJ events to occur with higher frequency during these two months (Fig. 4); however, the same signal is not apparent in SB-LLJ events.
Past studies have shown evidence of soil moisture feedback to precipitation from wet soils (Brimelow et al. 2011) and dry soils (Westra et al. 2012). Mechanistically, wet soils partition incoming energy into a higher ratio of latent heat over sensible heat, which increases moist static energy in the near-surface atmosphere (Pal and Eltahir 2001). Higher atmospheric humidity near the surface lowers the lifting condensation level and increases CAPE (Taylor and Lebel 1998). Drier-than-normal soils lead to increases in sensible heat flux and rapid PBL growth leading to entrainment of drier, warmer air (Ek and Holtslag 2004). Both result in increased air temperatures and decreased atmospheric humidity near the surface over dry soils. Despite less moisture flux from drier soils to the atmosphere, convection may be favored over dry soils because of rapid PBL growth, particularly if the PBL reaches the lifting condensation level (Santanello et al. 2011). Our results are consistent with both mechanisms for soil moisture feedback over wet and dry soils, in that events over wetter-than-normal soils correspond with increased humidity and stronger CAPE values. Events over drier-than-normal soils coincided with increased air temperature and decreased atmospheric humidity. The potential dominant soil moisture–precipitation feedback mechanism appears to be dependent on the presence or absence of the LLJ.
Although analyses presented here are consistent with positive and negative soil moisture feedbacks, this study does not confirm a causal mechanism linking soil moisture to subsequent precipitation. Interestingly, there is no statistically significant relationship between soil moisture and precipitation if we combine all of the convective events examined in this study into a single class (results not shown). This agrees with Taylor et al. (2012), who did not find any statistically significant positive or negative soil moisture–precipitation feedbacks in the U.S. Great Plains.
Despite the advantages of using in situ soil moisture observations, the soil moisture data are also a limitation of this study because of the measurement time (0600 UTC), as in situ observations later in the morning of precipitation would better characterize the land surface prior to convection. The spatial representativeness of in situ point soil moisture and the spatial density of soil moisture observations are also limitations of the Oklahoma Mesonet dataset. Although the Oklahoma Mesonet is the densest mesoscale soil moisture observation network in the United States, the average station spacing is still greater than 75 km per station. Station sparseness is the primary factor for the lack of in situ observation-based studies in soil moisture–precipitation coupling literature. In this case, the relatively large spacing between observing stations may have impacted the spatial variability of soil moisture and possibly diluted any smaller-scale preferences for rain to fall over soils that are drier or wetter with respect to adjacent soils (e.g., Taylor et al. 2012).
Other limitations include the CMORPH dataset temporal and spatial resolution, as a higher (temporal and spatial) resolution precipitation dataset would also allow us to better discern between events attributable to large-scale systems (e.g., frontal activity) and those associated with mesoscale convective systems. Future studies would be improved by the use of higher spatial and temporal resolution precipitation data such as the National Weather Service Stage IV radar product. In addition, this study focused on afternoon precipitation; however, the precipitation maximum in the southern Great Plains often occurs at night. We did not exclude afternoon precipitation events if rainfall occurred during the previous evening. Although nighttime-to-daytime precipitation persistence could influence soil moisture–precipitation coupling, it is beyond the scope of the present research.
5. Conclusions
This study used in situ soil moisture data to determine if afternoon precipitation occurs preferentially over drier or wetter soils in Oklahoma. We separated event days based on synoptic conditions and the presence of the Great Plains LLJ to isolate conditions representative of large-scale atmospheric forcing from those without. Events occurring during the presence (absence) of the LLJ fell preferentially over drier (wetter)-than-normal soils. These preferences were statistically significant for all event classes (at 95% confidence level). Soil moisture on the morning of events had statistically significant correlations with both noontime near-surface atmospheric humidity and temperature, such that humidity (temperature) increased (decreased) as soil moisture increased. Precipitation events with an LLJ present occurred more frequently and with larger CAPE anomalies when low-level air temperature was warmer than normal.
Our results show statistically significant preferences for afternoon precipitation to fall over wet soils and dry soils, depending on the presence or absence of the LLJ. This suggests that the Great Plains LLJ exerts not only a large dynamic influence on the convective environment of the southern United States, but may also impact land–atmosphere interactions that could potentially manifest through different mechanisms when the jet is present or absent.
Acknowledgments
We gratefully acknowledge the National Science Foundation (Award AGS-1056796) for funding this work. We would also like to thank the Oklahoma Mesonet for providing soil moisture observations used in this work. Soil moisture from the Mesonet can be obtained online at www.mesonet.org/index.php/weather/category/past_data_files.
REFERENCES
Alfieri, L., Claps P. , D’Odorico P. , Laio F. , and Over T. M. , 2008: An analysis of the soil moisture feedback on convective and stratiform precipitation. J. Hydrometeor., 9, 280–291, doi:10.1175/2007JHM863.1.
Basara, J. B., and Crawford K. C. , 2002: Linear relationships between root-zone soil moisture and atmospheric processes in the PBL. J. Geophys. Res., 107, 4274, doi:10.1029/2001JD000633.
Bonner, W. D., 1968: Climatology of the low level jet. Mon. Wea. Rev., 96, 833–850, doi:10.1175/1520-0493(1968)096<0833:COTLLJ>2.0.CO;2.
Brimelow, J. C., Hanesiak J. M. , and Burrows W. R. , 2011: Impacts of land–atmosphere feedbacks on deep, moist convection on the Canadian Prairies. Earth Interact., 15, doi:10.1175/2011EI407.1.
Crow, W. T., Drusch M. , and Wood E. F. , 2001: An observation system simulation experiment for the impact of land surface heterogeneity on AMSR-E soil moisture retrieval. IEEE Trans. Geosci. Remote Sens., 39, 1622–1631, doi:10.1109/36.942540.
Dirmeyer, P. A., Zeng F. J. , Ducharne A. , Morrill J. C. , and Koster R. D. , 2000: The sensitivity of surface fluxes to soil water content in three land surface schemes. J. Hydrometeor., 1, 121–134, doi:10.1175/1525-7541(2000)001<0121:TSOSFT>2.0.CO;2.
Dirmeyer, P. A., Schlosser C. A. , and Brubaker K. L. , 2009: Precipitation, recycling, and land memory: An integrated analysis. J. Hydrometeor., 10, 278–288, doi:10.1175/2008JHM1016.1.
Ek, M., and Holtslag A. , 2004: Influence of soil moisture on boundary layer cloud development. J. Hydrometeor., 5, 86–99, doi:10.1175/1525-7541(2004)005<0086:IOSMOB>2.0.CO;2.
Ferguson, C. R., and Wood E. F. , 2011: Observed land–atmosphere coupling from satellite remote sensing and reanalysis. J. Hydrometeor., 12, 1221–1254, doi:10.1175/2011JHM1380.1.
Findell, K. L., and Eltahir E. A. B. , 2003: Atmospheric controls on soil moisture–boundary layer interactions. Part I: Framework development. J. Hydrometeor., 4, 552–569, doi:10.1175/1525-7541(2003)004<0552:ACOSML>2.0.CO;2.
Findell, K. L., Gentine P. , Lintner B. R. , and Kerr C. , 2011: Probability of afternoon precipitation in eastern United States and Mexico enhanced by high evaporation. Nat. Geosci., 4, 434–439, doi:10.1038/ngeo1174.
Fischer, E. M., Seneviratne S. I. , Lüthi D. , and Schär C. , 2007: Contribution of land–atmosphere coupling to recent European summer heat waves. Geophys. Res. Lett., 34, L06707, doi:10.1029/2006GL029068.
Ford, T. W., and Quiring S. M. , 2014: Comparison and application of multiple methods for temporal interpolation of daily soil moisture. Int. J. Climatol., 34, 2604–2621, doi:10.1002/joc.3862.
Ford, T. W., Harris E. , and Quiring S. M. , 2014: Estimating root zone soil moisture using near-surface observations from SMOS. Hydrol. Earth Syst. Sci., 18, 139–154, doi:10.5194/hess-18-139-2014.
Frye, J. D., and Mote T. L. , 2010: Convection initiation along soil moisture boundaries in the southern Great Plains. Mon. Wea. Rev., 138, 1140–1151, doi:10.1175/2009MWR2865.1.
Garcia-Carreras, L., Parker D. J. , and Marsham J. H. , 2011: What is the mechanism for the modification of convective cloud distributions by land surface–induced flows? J. Atmos. Sci., 68, 619–634, doi:10.1175/2010JAS3604.1.
Gu, L., and Coauthors, 2006: Direct and indirect effects of atmospheric conditions and soil moisture on surface energy partitioning revealed by a prolonged drought at a temperate forest site. J. Geophys. Res., 111, D16102, doi:10.1029/2006JD007161.
Higgins, R., Yao Y. , Yarosh E. , Janowiak J. E. , and Mo K. , 1997: Influence of the Great Plains low-level jet on summertime precipitation and moisture transport over the central United States. J. Climate, 10, 481–507, doi:10.1175/1520-0442(1997)010<0481:IOTGPL>2.0.CO;2.
Illston, B. G., Basara J. B. , Fiebrich C. A. , Crawford K. C. , Hunt E. , Fisher D. K. , Elliott R. , and Humes K. , 2008: Mesoscale monitoring of soil moisture across a statewide network. J. Atmos. Oceanic Technol., 25, 167–182, doi:10.1175/2007JTECHA993.1.
Jackson, T. J., Cosh M. H. , Bindlish R. , and Du J. , 2007: Validation of AMSR-E soil moisture algorithms with ground based networks. IEEE Int. Geoscience and Remote Sensing Symp. 2007, Barcelona, Spain, IEEE, 1181–1184, doi:10.1109/IGARSS.2007.4423015.
Jackson, T. J., and Coauthors, 2010: Validation of Advanced Microwave Scanning Radiometer soil moisture products. IEEE Trans. Geosci. Remote Sens., 48, 4256–4272, doi:10.1109/TGRS.2010.2051035.
Jones, A. R., and Brunsell N. A. , 2009: Energy balance partitioning and net radiation controls on soil moisture–precipitation feedbacks. Earth Interact., 13, doi:10.1175/2009EI270.1.
Joyce, R. J., Janowiak J. E. , Arkin P. A. , and Xie P. , 2004: CMORPH: A method that produces global precipitation estimates from passive microwave and infrared data at high spatial and temporal resolution. J. Hydrometeor., 5, 487–503, doi:10.1175/1525-7541(2004)005<0487:CAMTPG>2.0.CO;2.
Kang, S., and Bryan G. H. , 2011: A large-eddy simulation study of moist convection initiation over heterogeneous surface fluxes. Mon. Wea. Rev., 139, 2901–2917, doi:10.1175/MWR-D-10-05037.1.
Koster, R. D., Suarez M. J. , Higgins R. W. , and Van den Dool H. M. , 2003: Observational evidence that soil moisture variations affect precipitation. Geophys. Res. Lett., 30, 1241, doi:10.1029/2002GL016571.
Koster, R. D., and Coauthors, 2004: Regions of strong coupling between soil moisture and precipitation. Science, 305, 1138–1140, doi:10.1126/science.1100217.
Koster, R. D., and Coauthors, 2011: The second phase of the Global Land–Atmosphere Coupling Experiment: Soil moisture contributions to subseasonal forecast skill. J. Hydrometeor., 12, 805–822, doi:10.1175/2011JHM1365.1.
Matyas, C. J., and Carleton A. M. , 2010: Surface radar-derived convective rainfall associations with Midwest US land surface conditions in summer seasons 1999 and 2000. Theor. Appl. Climatol., 99, 315–330, doi:10.1007/s00704-009-0144-7.
McPherson, R. A., 2007: A review of vegetation—Atmosphere interactions and their influences on mesoscale phenomena. Prog. Phys. Geogr., 31, 261–285, doi:10.1177/0309133307079055.
McPherson, R. A., Stensrud D. J. , and Crawford K. C. , 2004: The impact of Oklahoma’s winter wheat belt on the mesoscale environment. Mon. Wea. Rev., 132, 405–421, doi:10.1175/1520-0493(2004)132<0405:TIOOWW>2.0.CO;2.
Meng, L., and Quiring S. M. , 2010a: Examining the influence of spring soil moisture anomalies on summer precipitation in the U.S. Great Plains using the Community Atmosphere Model version 3. J. Geophys. Res., 115, D21118, doi:10.1029/2010JD014449.
Meng, L., and Quiring S. M. , 2010b: Observational relationship of sea surface temperatures and precedent soil moisture with summer precipitation in the U.S. Great Plains. Int. J. Climatol., 30, 884–893, doi:10.1002/joc.1941.
Mesinger, F., and Coauthors, 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87, 343–360, doi:10.1175/BAMS-87-3-343.
Myoung, B., and Nielsen-Gammon J. W. , 2010: 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, 4461–4473, doi:10.1175/2010JCLI2946.1.
Pal, J. S., and Eltahir E. A. B. , 2001: Pathways relating soil moisture conditions to future summer rainfall within a model of the land–atmosphere system. J. Climate, 14, 1227–1242, doi:10.1175/1520-0442(2001)014<1227:PRSMCT>2.0.CO;2.
Pal, J. S., and Eltahir E. A. B. , 2003: A feedback mechanism between soil-moisture distribution and storm tracks. Quart. J. Roy. Meteor. Soc., 129, 2279–2297, doi:10.1256/qj.01.201.
Pielke, R. A., 2001: Influence of the spatial distribution of vegetation and soils on the prediction of cumulus convective rainfall. Rev. Geophys., 39, 151–177, doi:10.1029/1999RG000072.
Robock, A., Vinnikov K. Y. , Srinivasan G. , Entin J. K. , Hollinger S. E. , Speranskaya N. A. , Liu S. , and Namkhai A. , 2000: The global soil moisture data bank. Bull. Amer. Meteor. Soc., 81, 1281–1300, doi:10.1175/1520-0477(2000)081<1281:TGSMDB>2.3.CO;2.
Roundy, J. K., Ferguson C. R. , and Wood E. F. , 2013: Temporal variability of land–atmosphere coupling and its implications for drought over the Southeast United States. J. Hydrometeor., 14, 622–635, doi:10.1175/JHM-D-12-090.1.
Santanello, J. A., and Peters-Lidard C. D. , 2013: Diagnosing the nature of land–atmosphere coupling: A case study of dry/wet extremes in the U.S. southern Great Plains. J. Hydrometeor., 14, 3–24, doi:10.1175/JHM-D-12-023.1.
Santanello, J. A., Peters-Lidard C. D. , and Kumar S. V. , 2011: Diagnosing the sensitivity of local land–atmosphere coupling via the soil moisture–boundary layer interaction. J. Hydrometeor., 12, 766–786, doi:10.1175/JHM-D-10-05014.1.
Scott, B. L., Ochsner T. E. , Illston B. G. , Fiebrich C. A. , Basara J. B. , and Sutherland A. J. , 2013: New soil property database improves Oklahoma Mesonet soil moisture estimates. J. Atmos. Oceanic Technol., 30, 2585–2595, doi:10.1175/JTECH-D-13-00084.1.
Seneviratne, S. I., Corti T. , Davin E. L. , Hirschi M. , Jaeger E. B. , Lehner I. , Orlowsky B. , and Teuling A. J. , 2010: Investigating soil moisture–climate interactions in a changing climate: A review. Earth-Sci. Rev., 99, 125–161, doi:10.1016/j.earscirev.2010.02.004.
Taylor, C. M., and Lebel T. , 1998: Observational evidence of persistent convective-scale rainfall patterns. Mon. Wea. Rev., 126, 1597–1607, doi:10.1175/1520-0493(1998)126<1597:OEOPCS>2.0.CO;2.
Taylor, C. M., and Ellis R. J. , 2006: Satellite detection of soil moisture impacts on convection at the mesoscale. Geophys. Res. Lett., 33, L03404, doi:10.1029/2005GL025252.
Taylor, C. M., Parker D. J. , and Harris P. P. , 2007: An observational case study of mesoscale atmospheric circulations induced by soil moisture. Geophys. Res. Lett., 34, L15801, doi:10.1029/2007GL030572.
Taylor, C. M., Gounou A. , Guichard F. , Harris P. P. , Ellis R. J. , Couvreux F. , and De Kauwe M. , 2011: Frequency of Sahelian storm initiation enhanced over mesoscale soil-moisture patterns. Nat. Geosci., 4, 430–433, doi:10.1038/ngeo1173.
Taylor, C. M., de Jeu R. A. , Guichard F. , Harris P. P. , and Dorigo W. A. , 2012: Afternoon rain more likely over drier soils. Nature, 489, 423–426, doi:10.1038/nature11377.
Teuling, A. J., and Seneviratne S. I. , 2008: Contrasting spectral changes limit albedo impact on land–atmosphere coupling during the 2003 European heat wave. Geophys. Res. Lett., 35, L03401, doi:10.1029/2007GL032778.
Wagner, W., Naeimi V. , Scipal K. , de Jeu R. , and Martínez-Fernández J. , 2007: Soil moisture from operational meteorological satellites. Hydrogeol. J., 15, 121–131, doi:10.1007/s10040-006-0104-6.
Wei, J., and Dirmeyer P. A. , 2012: Dissecting soil moisture–precipitation coupling. Geophys. Res. Lett., 39, L19711, doi:10.1029/2012GL053038.
Wei, J., Dickinson R. E. , and Chen H. , 2008: A negative soil moisture–precipitation relationship and its causes. J. Hydrometeor., 9, 1364–1376, doi:10.1175/2008JHM955.1.
Westra, D., Steeneveld G. J. , and Holtslag A. A. M. , 2012: Some observational evidence for dry soils supporting enhanced relative humidity at the convective boundary layer top. J. Hydrometeor., 13, 1347–1358, doi:10.1175/JHM-D-11-0136.1.
Whiteman, C. D., Bian X. , and Zhong S. , 1997: Low-level jet climatology from enhanced rawinsonde observations at a site in the southern Great Plains. J. Appl. Meteor., 36, 1363–1376, doi:10.1175/1520-0450(1997)036<1363:LLJCFE>2.0.CO;2.
Wu, W., and Dickinson R. E. , 2004: Time scales of layered soil moisture memory in the context of land–atmosphere interaction. J. Climate, 17, 2752–2764, doi:10.1175/1520-0442(2004)017<2752:TSOLSM>2.0.CO;2.
Wu, Y., and Raman S. , 1998: The summertime Great Plains low level jet and the effect of its origin on moisture transport. Bound.-Layer Meteor., 88, 445–466, doi:10.1023/A:1001518302649.
Xia, Y., Sheffield J. , Ek M. B. , Dong J. , Chaney N. , Wei H. , Meng J. , and Wood E. F. , 2014: Evaluation of multi-model simulated soil moisture in NLDAS-2. J. Hydrol., 512, 107–125, doi:10.1016/j.jhydrol.2014.02.027.