Baffaut - Multi-Scale Monitoring
0 likes415 views
This document summarizes research from the Goodwater Creek Experimental Watershed on the effects of no-till and cover crops on water quality. The summary is: 1) No-till and cover crops reduced sediment loss but did not change nitrate transport and increased dissolved phosphorus transport. 2) Multiple factors like land use change, urbanization, and stream processes influence watershed water quality and make it difficult to detect trends or impacts from individual changes. 3) To better detect water quality impacts from management changes, the changes need to be implemented quickly, targeted spatially, of large magnitude, and address the key processes causing water quality issues.
Baffaut - Multi-Scale Monitoring
- 1. Claire Baffaut Nutrient Management and Edge of Field monitoring: from the Great Lakes to the Gulf Memphis, TN 2015 Multi-scale Monitoring for Improved Nutrient Management
- 2. Challenges of tracking progress in streams Additional processes Lack of targeting Lack of control: Conflicting factors Lack of spatial and temporal information No replication Data interpretation is difficult. Changes are small and difficult to detect.
- 3. Goodwater Creek Experimental Watershed
- 4. • Stream bank erosion • Subsurface and ground water contributions. • Nutrient storage in and resuspension from the streambed sediment. • Filtering and nutrient uptake by riparian buffers. Multiple processes Bank sediment 87% Overland sediment 13% Instream sediment in Otter and Crooked Creek Bank nitrogen 23% Overland nitrogen 77% Instream nitrogen in Otter and Crooked Creek Willett et al., JSWC 2012.
- 5. Replication Relatively easy at the plot scale: - Same soils - Same slope - Same initial conditions - Same weather - Same size and shape - Same orientation Repeated treatments on several plots allow statistical analysis to detect significant differences
- 6. Other monitoring and analysis strategies • Before and after analysis: difficult because change in land agricultural practices is gradual. • Trend analysis. • Multiple regression analysis. • Multiple scale monitoring.
- 7. BMPs in Goodwater Creek Experimental Watershed 15% of the watershed in 17 years !
- 8. • Before and after analysis. • Trend analysis: – Issues with conflicting factors. – Effect of a strong random component • Multiple regression analysis. • Multiple scale monitoring. Other monitoring and analysis strategies
- 9. Trend Analysis of Flow in GCEW Year peak flow : 8 mm more per decade Number of flooded days: 2 more days per decade
- 10. Agricultural Land use 1967 Row crop (corn and soybean) Small grain (wheat) Hay and pasture 2006
- 11. Soil erosion and water storage capacity • 13 cm (5.1 in) in 150 years • 3.5 cm (1.4 in) in 40 years • 14% of water storage capacity. Top soil loss (cm) -45 -20 0 20 45 (Lerch et al. 2005, JSWC)
- 12. 1967 2009 Tributary crossing in Centralia Urbanization • 11% more people • 71% more houses
- 13. Conservation Practices 0 2 4 6 8 10 12 14 16 18 0 50 100 150 200 250 %watershedprotected Areaprotected(hectares) Grassed waterway Terraces Seeding Sod busting Grazing system Lagoon Filter Strip Diversion CP33,CRP Buffer Cumulative % land protected
- 14. • Before and after analysis. • Trend analysis. • Multiple regression analysis: requires good spatial and temporal knowledge of what is happening in the watershed. • Multiple scale monitoring. Other monitoring and analysis strategies
- 15. Multiple regression analysis • Requires good spatial and temporal information of: – Weather – Land use – Crop distribution – Land management, including • Cropland management and best management practices • Sanitary sewage treatment • Management of urban areas • Management of pastures
- 16. Multiple regression • No trend of nitrate loads over 92-06 in GCEW (O’Donnell, 2010). • Decreasing trend over 1992-2010 (Lerch et al., 2015), possibly linked to decrease in wheat production. • No BMP linked variable found significant. • Not the right BMPs? • Not the right location? Crop land Pasture & grass Impervious areas Critical areas Conservation practices
- 17. Time needed to detect change • Mean Square Error of model was used to estimate the monitoring period needed to detect a future change Predicted number of years needed to detect load reduction Nitrate load reduction Season 5% 10% 20% 25% Year 92 24 7 4 Spring 185 49 13 9 O’Donnell, 2010
- 18. • Before and after analysis. • Trend analysis. • Multiple regression analysis: requires good spatial and temporal knowledge of what is happening in the watershed. • Multiple scale monitoring. Other monitoring and analysis strategies
- 19. Multiple scale monitoring Stream Weir W1 Field 1993-2002 Mulch tilled corn-soybean 2004-2014 Precision Agricultural System (PAS)
- 20. Field 1 Pre-PAS 1993-2002 PAS 2004-2013 South 40 acres North 52 acres Odd year Corn (sorghum in 95) N: pre-plant UAN, incorporated P: 1993, 1995, 2001 incorporated Cultivation Corn / Cover Crop N: at planting + top dress early summer No-till Wheat / Cover Crop N: top dress in April No-till Even year Soybean Cultivation Soybean / Cover Crop P: 2004, 2006, 2008 broadcast No-till Soybean / wheat N: at wheat planting P: 2004, 2006, 2008 broadcast No-till
- 21. Effect of no-till and cover crops Flow
- 22. Effect of no-till and cover crop Dissolved P loss
- 23. Effect of no-till and cover crop Nitrate-N loss
- 24. Effect of no-till and cover crop Sediment
- 26. Summary • No-till and cover crops did: Reduce sediment Did not change Nitrate-N transport Did increase Dissolved P transport • Agronomic practices, land use change, urbanization, stream processes and climate all contribute to modifying the runoff/sediment/nutrient yield regime of a watershed and make it difficult to: detect a trend, discern whether detected trends are due to any one factor.
- 27. Implications To improve detection of water quality trends resulting from management changes, these changes should be: • Implemented within a short time. • Spatially targeted. • Of large magnitude. • Addressing the processes that cause the problem documented by the monitoring. Scale matters!
- 28. Data Management • Equipment fails data gaps fill in the gaps • Sediment loss measurements • Inaccurate data – Flow > Precipitation – Issues with small events • Meta data
- 29. Acknowledgements Cropping Systems and Water Quality Research Unit Newell Kitchen, Ken Sudduth, Bob Lerch Matt Volkmann, Kurt Holiman, Mark Olson, Aaron Beshears, Teri Oster, Scott Drummond, Bettina Coggeshall. ARS CEAP LTAR