On the use of modelling, observations and remote sensing to better understand the Canadian prairie soil-crop-atmosphere system
Brimelow, Julian Charles
MetadataShow full item record
Thunderstorms have been identified as an important component of the hydrological cycle on the Canadian Prairies, a region that is postulated to have the potential to exert a detectable influence on convective precipitation in the summer. However, very little work has been undertaken exploring and elucidating those aspects of biophysical forcing on the Canadian Prairies that affect lightning activity during the summer months, the constraints under which any linkages operate, and the mechanisms by which surface anomalies modify the structure and moisture content of the convective boundary layer (CBL) so as to modulate lightning activity. Evapotranspiration (ET) from the soil and vegetation canopy is known to be important for modulating the moisture content in the CBL, and this in turn has important implications for the initiation and intensity of deep, moist convection. The Second Generation Prairie Agrometeorological Model (PAMII) of Raddatz (1993) has been used extensively for the purpose of quantifying the evolution of soil moisture and ET in response to atmospheric drivers on the Canadian Prairies. However, the ability of PAMII to simulate the evolution of root-zone soil moisture and ET during the growing season has yet to be verified against a comprehensive set of in-situ observations. In this thesis, we address the above knowledge gaps using unique datasets comprising observed lightning flash data, satellite-derived Normalized Difference Vegetation Index (NDVI) data, observed atmospheric soundings, in-situ soil moisture observations and estimates of daily ET from eddy-covariance systems. A thorough quantitative validation of simulations of root-zone soil moisture and ET from PAMII was undertaken against in-situ soil moisture measurements and ET from eddy-covariance systems at sites on the Canadian Prairies. Our analysis demonstrates that PAMII shows skill in simulating the evolution of bulk root-zone soil moisture content and ET during the growing season, and for contrasting summer conditions (i.e., wet versus dry). As part of the soil moisture validation, a novel multi-model pedotransfer function ensemble technique was developed to quantify the uncertainty in soil moisture simulations arising from errors in the specified soil texture and associated soil hydraulic properties. An innovative approach was used to explore linkages between the terrestrial surface and deep, moist convection on the Canadian Prairies, using datasets which avoid many of the problems encountered when studying linkages between soil moisture and thunderstorm activity. This was achieved using lightning flash data in unison with remotely sensed NDVI data. Specifically, statistical analysis of the data over 38 Census Agricultural Regions (CARs) on the Canadian Prairies for 10 summers from 1999 to 2008 provided evidence for a surface-convection feedback on the Canadian Prairies, in which drought tends to perpetuate drought with respect to deep, moist convection. The constraints in which such a feedback operates (e.g., areal extent and magnitude of the NDVI anomalies) were also identified. For example, our data suggest that NDVI anomalies and lightning duration are asymmetric, with the relationship between NDVI and lightning duration strengthening as the area and amplitude of the negative NDVI anomaly (less vegetation vigour) increases. Finally, we focused on how surface anomalies over the Canadian Prairies can condition the CBL so as to inhibit or facilitate thunderstorm activity, while also considering the role of synoptic-scale forcing on modulating summer thunderstorm activity. We focused on a CAR located over central Alberta for which observed lightning flash data, NDVI data, and in-situ sounding data were available for 11 summers from 1999 to 2009. Our analysis suggests that storms over this region are more likely to develop and are longer-lived or more widespread when they develop in an environment in which the surface and upper-air synoptic-scale forcings are synchronized. On days when a surface or upper-air feature is present, storms are more likely to be triggered when NDVI is much above average, compared to when NDVI is much below average. We propose a conceptual model, based almost entirely on observations, which integrates our findings to describe how a reduction in vegetation vigour modulates the partitioning of available energy into sensible and latent heat fluxes at the surface, thereby modulating the lifting condensation level heights, which in turn affect lightning duration.