Paleoecological investigations of 28 lakes and bogs provide the basis for a Holocene paleohydrological record for central Alberta. Shallow basins, empty during the early Holocene, began flooding shortly after 8000 years ago; most filled from 6500 to 4500 years ago, and none filled later than 3000 years ago. Pollen of a hypersaline indicator genus, Ruppia, was found in cores of lakes where the plant does not presently grow, indicating lower lake levels, evaporation stress, and increased salinity from 8000 to 3000 years ago. Moore Lake (54°30′N, 100°30′W) may have dropped 15 m from 9200 to 5900 years ago, and Lofty Lake (54°44′N, 112°29′W) largely dried up between 8700 and 6300 years ago. Lower water levels are suggested for several lakes by shifts in fossil diatom populations from planktonic to benthic, particularly epipelic species including pioneering taxa. These shallow, warm lakes were highly productive and sedimentary pigment levels reached maximum values between about 9000 and 4000 years ago.The paleohydrological record indicates early Holocene aridity, with the onset and development of moister conditions between 8000 and 3000 years ago, by which time modern climatic and vegetation conditions had been established. This record fits the Holocene climate predicted by the general-circulation-model simulations based on orbital perturbations.

No abstract

Herbicide concentration in runoff varies dramatically within a storm, therefore storm pattern is postulated to have a significant impact on herbicide loss. We evaluated the effects of storm pattern and soil moisture content on herbicide loss in runoff, and used our data to validate the uniform mixing concept for modeling herbicide transfer to runoff. Atrazine and metolachlor were surface applied to air‐dried soil at rates of 1.12 and 2.24 kg/ha, respectively. Two soils (Cecil sandy loam and Miami silty loam), four storm patterns (uniform, advanced peak, intermediate peak, and delayed peak), and two moisture levels (wet and dry) were used. Dissolved herbicide losses from the advanced peak were twice those from the other patterns for both herbicides for Cecil soil. The significance of rain pattern effects was reduced for Miami soil, but metolachlor loss from the advanced peak was 1.9 times that from the delayed peak. Rain pattern affected sediment‐bound herbicide loss, but no consistent trend was shown for both soils. Under wet conditions a greater potential for dissolved and sediment‐bound herbicide losses was shown for both herbicides on Miami soil than on Cecil soil. Measured herbicide concentrations in runoff decreased exponentially with cumulative rainfall depth, and were adequately described by the uniform mixing model. Results show that information on intensity distribution and transient infiltration rate is needed to better predict both instantaneous and total herbicide losses during a storm.

No abstract