Drought, Climate Variability and Water Resources in Western Canada J.W. Pomeroy, K.R. Shook, R.N. Armstrong, X. Fang Centre for Hydrology University of Saskatchewan, Saskatoon, Canada www.usask.ca/hydrology john.pomeroy@usask.ca
Canadian Rocky Mountains are the Headwaters Rocky Mountains for much of Canadian Prairies North America
Canadian Rocky Mountains Cold alpine and sub-alpine climate
Rivers Draining the Rocky Mountains Supply Prairie Cities and Irrigation
Canadian Prairie Agriculture: mostly dryland grain farming & pasture Cold, continental, semi-arid climate Precipitation on average 350 mm Grain Growing 125 mm soil water reserves needed from snowmelt 175 mm spring rainfall needed Roughly 300 kg/ha increased wheat yield for each extra 25 mm of water added
Canadian Prairie Drought of 1999-2004 Most Expensive Natural Disaster in Canadian History $5.8 billion decline in GDP 2001-2002 $3.6 billion drop in agricultural production, 2001-2002 41,000 jobs lost British Columbia & Alberta forest fires Saskatchewan dust storms
OBJECTIVE OF DRI the Drought Research Initiative To better understand the physical characteristics of and processes influencing Canadian Prairie droughts, and to contribute to their better prediction, through a focus on the recent severe drought that began in 1999 www.drinetwork.ca
DRI THEMES 1. Quantify the physical features, flows of water and energy into and out of the region, and storage and redistribution within the region 2. Improve the understanding of processes and feedbacks governing the formation, evolution, cessation and structure of the drought 3. Assess and reduce uncertainties in the prediction of drought 4. Compare the similarities and differences of current drought to previous droughts and those in other regions 5. Apply our progress to address critical issues of importance to society
1. QUANTIFY THE DROUGHT Observational Networks Surface Storage Change 200 m GRACE satellite Crestomere Lake Gull Lake Ponoka Sylvan Lake Red Deer Dickson DamInnisfail Olds Pine Lake Elnora Buffalo Lake Buffalo Lake Stettler Kirkpatrick Lake Duck Lake Hague Warman Saskatoon Saskatoon Vanscoy Blucher Bruno Many Springs Irricana Calgary Carseland OkotoksHigh River DrumhellerHand Hills Sounding Creek Sibbald Big Stone Gem Cessford Cluny Buffalo North Duchess Cavendish Tyner Swanson Conquest Enchant Barons Medicine Hat Mud Lake Orton Lethbridge Oldman Dam Forty Mile Coulee Elkwater Warner Pakowki Cypress Verlo Instow Shaunavon Swift Current Smith Coulee Del Bonita Legend Wells in South Saskatchewan Observation Well Location 0 15 30 60 90 120 Kilometers
2. UNDERSTAND THE DROUGHT Drought Vertical Scale Drought Non-drought Non-drought Storage of Water Horizontal Flux of Water
3. SIMULATE AND PREDICT THE Global Climate model DROUGHT NWP Model AGCM3 200 km Global Reanalysis GEM 15 km Regional Climate Model CRCM4 45 km Cloud-Resolving Model GEM-LAM 2.5 km Forcings and Initial conditions Process Parameterizations LDAS data flow LDAS Land Surface Hydrology Models CLASS, MESH 15 km CPM, CRHM 1 km Hydrological Process Models Regional Analysis & obs (for LDAS)
Canadian Prairie Runoff Generation Snow Redistribution to Channels Spring melt and runoff Dry non-contributing areas to runoff Water Storage in Wetlands
Runoff Non-Contributing Areas to Streamflow a Prairie Characteristic
Prairie Streamflow very seasonal Smith Creek, Saskatchewan Streamflow m 3 per second 25 20 15 10 5 0 Drainage area ~ 450 km 2 Average 1975-2006 1995 High Year 2000 Low Year No baseflow from groundwater 27-Dec 27-Nov 28-Oct 28-Sep 29-Aug 30-Jul 30-Jun 31-May 01-May 01-Apr 02-Mar 31-Jan 01-Jan Hydrological drought can be viewed as the absence of prairie runoff
Sparse Prairie Streamflow Network for Drought Analysis Burn et al., 2008 Inadequate coverage to characterise prairie runoff
How to Characterise Hydrological Drought for the Prairies? River flows reflect Rocky Mountain, but not Prairie hydrological conditions. Streamflow network is very sparse. Modelling required to create a hydrological drought surface for the Prairies over small first order basins.
Cold Regions Hydrological Model (CRHM) Drought Hydrology Simulation Create virtual representative basin models, which can produce simple exceedence probability drought indices Allows comparison of basin response to drought conditions and to historical variability Requires high resolution driving data over entire Prairies for normal (1961-1990), non-drought (1962-1987) and drought periods (1999-2004) Snowfall, rainfall Temperature Humidity Wind speed
CRHM Prairie Module Structure Garnier and Ohmura s radiation module Global radiation Max. sunshine Observation module Temperature, Windspeed Relative Humidity Vapour Pressure Precipitation Sunshine hour module Sunshine hour Walmsley s windflow module Long-wave module Interception module Adjusted windspeed Rainfall Snowfall PBSM SWE Gray and Landine s Albedo module Adjusted short- and long-wave radiation Canopy effect adjustment for radiation module albedo EBSM snowmelt All-wave radiation module Short- and long-wave radiation Gray s snowmelt infiltration module, Green-Ampt infiltration Snow INF Rain INF Runoff Granger s evaporation module, Priestley and Taylor evaporation module Evap in recharge and rooting zones Wetland module Runoff Muskingum routing module
Impact of Wetland Change on Streamflow in a Wet Year Scenarios of Smith Creek Spring Discharge near Marchwell Daily Mean Discharge (m 3 /s) 30 "Normal Condition" 25 high natural wetland extent 20 minimum wetland extent 15 10 5 0 01-Mar 10-Mar 19-Mar 28-Mar 06-Apr 15-Apr 24-Apr 03-May 12-May 21-May 30-May 1995 1995 Flood: Record High Discharge Volume
Impact of Wetland Change on Streamflow in a Drought Year Scenarios of Smith Creek Spring Discharge near Marchwell Daily Mean Discharge (m 3 /s) 1.6 1.4 "Normal Condition" 1.2 high natural wetland extent 1 minimum wetland extent 0.8 0.6 0.4 0.2 0 11-Feb 20-Feb 29-Feb 09-Mar 18-Mar 27-Mar 05-Apr 14-Apr 23-Apr 02-May 2000 2000 Drought: Lowest Discharge Volume on Record
Drought Hydrology Simulations Station locations & Prairie ecozone
Summer Growing Season Climate Normals (1971-2000) Dry southwestern zone is suited for pasture rather than dryland farming
CRHM Surface Water Drought Modelling CRHM was used to create virtual model of typical prairie upland basin Model was run over climate normal period (1961-1990) 1990) Output during drought period was compared to normal period and spatially interpolated
CRHM model of small upland Small stream basin prairie stream basin CRHM model HRU1 HRU2 HRUs 1 and 2 alternate between cropped and fallow HRU3 is grassed HRU3
Prairie Spring Discharge 2001 Drought
Prairie Spring Discharge 2005 Wet Year
Spatial Variation of Prairie Soil Moisture (Drought vs Wet) Mean for normal period 332 mm Drought period: distribution wide, variance large, median > mean Wetter period: distribution loses low soil moisture, variance smaller, median < mean Drought Wet Probability density of soil moisture
Spatial Variation of Evapotranspiration (Drought vs Wet) Mean for normal period 352 mm Drought period: distribution wide, variance large, median >> mean Wetter period: distribution symmetric, variance greatly reduced Drought Wet Probability density of evapotranspiration
Climate Change Impacts Virtual basin model also useful for estimating climate change impacts on Prairie hydrology Climate warming has a strong effect on cold season hydrological processes.
Prairie Streamflow & Climate Change First more, then less Three most reliable climate change scenarios suggest increases in annual prairie winter temperature and precipitation from the 1961-1990 average: 2050 +2.6 ºC and +11% 2080 +4.7 ºC and +15.5% Using these scenarios in the virtual upland basin results in a 24% rise in 2050 spring runoff, but a 37% drop by 2080, compared to conditions in the mid 1970s.
70 60 50 40 30 20 10 0 Prairie Climate Change Winter Snow Winter Snow Accumulation at Bad Lake, SK Normal SWE (Winter of 1974/75) SWE (Winter of 2049/50) SWE (Winter of 2079/80) SWE (mm) 01/10/1974 15/10/1974 29/10/1974 12/11/1974 26/11/1974 10/12/1974 24/12/1974 07/01/1975 21/01/1975 04/02/1975 18/02/1975 04/03/1975 18/03/1975 01/04/1975 15/04/1975 29/04/1975
70 60 50 40 30 20 10 0 Prairie Climate Change Spring Runoff Spring Runoff from Creighton Tributary at Bad Lake, SK Runoff (mm) 01/10/1974 15/10/1974 29/10/1974 12/11/1974 26/11/1974 10/12/1974 24/12/1974 07/01/1975 21/01/1975 04/02/1975 18/02/1975 04/03/1975 18/03/1975 01/04/1975 15/04/1975 29/04/1975 Normal Spring Runoff (Spring of 1975) Spring Runoff (Spring of 2050) Spring Runoff (Spring of 2080)
Conclusions Prairie streamflow is inherently unreliable in drought, Hydrological droughts migrate across the Prairies. Drought increases the variability of soil moisture, evapotranspiration and snowpack Future climate change likely to first increase then decrease spring snowmelt runoff from prairie drainages. Flooding Shortages If droughts also increase, this means greater use will need to be made of mountain-derived river water supplies for agriculture
Thanks! The gracious invitation of ISACS-3 and financial support of National Natural Science Foundation of China made this presentation possible.