What causes cooling water temperature gradients in a forested stream reach?

Transcription

1 What causes cooling water temperature gradients in a forested stream reach? Grace Garner 1 ; Iain A. Malcolm 2 ; David M. Hannah 1 ; Jonathan P. Sadler 1 1 School of Geography, Earth & Environmental Sciences, University of Birmingham, UK 2 Marine Scotland Science, Freshwater Laboratory, Faskally, Pitlochry, UK gxg627@bham.ac.uk

2 Research context Woodland associated with decreases in: diurnal variability, mean and maximum water temperature Daytime cooling gradients observed for reaches shaded by trees downstream of clear cuts and open landuse Research gaps: Existing Studies confounded by cool groundwater inputs No explicit conceptualisation of the processes driving cooling patterns in forested reaches without groundwater inputs Why is this research important? Riparian vegetation advocated as climate change mitigation strategy in E.U. and N. America Process based understanding informs future riparian strategies to maximise benefits at minimal cost.

3 Aim and Objectives AIM: To quantify and model spatio-temporal variability in stream temperature and heat fluxes for an upland reach that transitions from open moorland to semi-natural woodland OBJECTIVES: 1. To quantify the magnitude of instantaneous longitudinal temperature gradients within the reach 2. To understand how and why water temperature changes as it travels beneath the forest and attribute to underlying processes

4 Study area & sites Area: Girnock Burn, upland tributary of the Aberdeenshire Dee Semi-natural catchment Reach: Established 5 m study reach Upstream (25 km 2 ) landuse dominated by heather moorland Within the reach, landuse transitions from moorland to semi-natural forest

5 Instrumentation water temperature loggers 3 automatic weather stations (AWSs) incl. thermistors 211 hemispherical photographs- to scale radiation measurements

6 Methods Stream energy budget: Q n = Q a + Q * + Q e + Q h + Q bhf Where: Q * Q h Q e Q * Q h Q e Q n = net energy flux Q a = longitudinal advected heat flux Q * = net radiative flux Q e = latent heat flux Q h = sensible heat flux Q bhf = bed heat flux Q a Q bhf Q bhf

7 Methods How and why does water change as it moves through the reach? Energy flux model incorporating flow routing (i.e Lagrangian model) Model simulated change in temperature of water parcels released from upstream reach boundary at 1 m resolution & 15 min. intervals T +/- Q w initial n T w +/- Q n T w +/- Q n T w

8 gy Flux (MJm -2 Energy d -1 ) Flux (MJm -2 d -1 ) Tw ( C) Q (m -3 s -1 )/ V (ms Energy -1 ) Flux (MJm -2 d -1 ) Energy Flux (MJm -2 d -1 ) Ta ( C) Q (m -3 s -1 )/ V (ms -1 ) Results: study period (a) 25 Avg. Maximum Avg. Minimum Ta (b).5 Avg. Maximum Q Avg. Minimum Q Q V (c) (b) Jul 2 Jul 3 Jul 4 Jul 5 Jul 6 Jul 7 Jul Jul Avg. Maximum Q Avg. Minimum Q Qn K* L* Qe Qh Qbhf 1 Jul 2 Jul 3 Jul 4 Jul 5 Jul 6 Jul 7 Jul Jul Q V (d) Jul 2 Jul 3 Jul 4 Jul 5 Jul 6 Jul 7 Jul Jul 7-day study period during summer 13 1 Jul 2 Jul 3 Jul 4 Jul 5 Jul 6 Jul 7 Jul Jul Minimal (4.2 mm) rainfall Worst-case scenario of high energy gains & low flows Jul Jul (e) (d) 1. 1 Jul 2 Jul 3 Jul 4 Jul 5 Jul 6 Jul 7 Jul Jul (f) AWS Open AWS FDS Difference 1 Jul 2 Jul 3 Jul 4 Jul 5 Jul 6 Jul 7 Jul Jul 1 Jul 2 Jul 3 Jul 4 Jul 5 Jul 6 Jul 7 Jul Jul

9 Energy Flux (MJm -2 d -1 ) Energy Flux (MJm -2 d -1 ) Tw ( C) Tw ( C) Energy Flux (MJm -2 d -1 ) Energy Flux (MJm -2 d -1 ) Energy Flux (MJm -2 d -1 ) Energy Flux (MJm -2 d -1 ) Energy Flux (MJm -2 d -1 ) Ta ( C) Ta ( C) Q (m -3 s -1 )/ V (ms -1 ) Q (m -3 s -1 )/ V (ms -1 ) Energy Flux (MJm -2 d -1 ) 25 Avg. Avg. Minimum Minimum Avg. Avg. Minimum Minimum Q Q Results: observed hydrometeorological data Avg. Avg. Maximum Maximum Ta Ta Avg. Avg. Maximum Maximum Q Q Q V Q V (c) ( 5 5 Qn L* Qh K* Qe Qbhf Jul 1 Jul 2 Jul 2 Jul 3 Open Jul 3 Jul 4 Jul 4 Jul 5 site Jul 5 Jul 6 Jul 6 Jul 7 Jul 7 Jul Jul Jul 1 Jul 1 Jul 2 Jul 2 Jul 3 Forest Jul 3 Jul 4 Jul 4 Jul 5 Jul site 5 Jul 6 Jul 6 1 Jul 7 Jul 7 Jul Jul Jul 1 Jul 2 JulForest 3 4 Jul site 5 Jul 6 2 Jul 7 Jul Jul 1 5 (c) (c) 1 1 (d) (d) 1 1 (e) ( Qn K* Qn K* L* Qe L* Qe Qh Qh Qbhf Qbhf 1 Jul 1 Jul 2 Jul 2 Jul 3 Jul 3 Jul 4 Jul 4 Jul 5 Jul 5 Jul 6 Jul 6 Jul 7 Jul 7 Jul Jul Jul Jul 1 Jul 2 Jul 2 Jul 3 Jul 3 Jul 4 Jul 4 Jul 5 Jul 5 Jul 6 Jul 6 Jul 7 Jul 7 Jul Jul Jul 5 1 Jul 2 Jul 3 Jul 4 Jul 5 Jul 6 Jul 7 Jul Jul (e) (e) Jul 1 Jul 2 Jul 2 Jul 3 Jul 3 Jul 4 Jul 4 Jul 5 Jul 5 Jul 6 Jul 6 Jul 7 Jul 7 Jul Jul Jul (f) (f) AWS AWS Open Open AWS AWS FDS FDS Difference Difference Energy exchange highly spatially (& temporally) heterogeneous Energy exchange at water column-atmosphere interface dominated energy budget 1 Jul 1 Jul 2 Jul 2 Jul 3 Jul 3 Jul 4 Jul 4 Jul 5 Jul 5 Jul 6 Jul 6 Jul 7 Jul 7 Jul Jul Jul (miniscule bed heat flux)

10 Tw ( C) Jul 5 Results: observed spatio-temporal T w patterns 1 Jul 2 Jul 3 Jul 4 Jul 5 Jul 6 Jul 7 Jul Jul (f) AWS Open AWS FDS Difference T w at upstream reach boundary up to 2.5 C warmer than downstream boundary Minima occur synchronously. Downstream maxima lag upstream by up to 1 hour Jul 1 Jul 2 Jul 3 Jul 4 Jul 5 Jul 6 Jul 7 Jul Jul Instantaneous cooling Tw ( C) Instantaneous differences increase with distance downstream Instantaneous warming

11 Results: estimated spatio-temporal Q n patterns Canopy Density (%) Qn (MJm 2 d 1 ) (MJm Qn (a) (c) Qn( MJm 2 d 1 ) Day two 2 3 2/7 23: /7 1: 2/7 : Distance downstream (m) 2/7 6: 2/7 : 4 6 (b) Distance downstream (m) /7 23:45 7/7 Day six2 (d) 2 d 1 ) 6/7 6/7 1: 5/7 6/7 : 4/7 3/7 6/7 6: 2/7 1/7 6/7 : Distance downstream (m) Distance downstream (m) Longitudinal patterns of net energy flux correspond broadly to canopy density Net energy gains to the stream during daytime Net energy losses overnight So, how do cooling gradients develop during the daytime if the stream is gaining heat?

12 Results: T w model performance Tw ( C) /7 6: 1/7 7: 1/7 : 1/7 9: /7 6: 2/7 7: 2/7 : 2/7 9: /7 6: 3/7 7: 3/7 : 3/7 9: Distance downstream (m) 4 6 4/7 6: 4/7 7: 4/7 : 4/7 9: 4 6 5/7 6: 5/7 7: 5/7 : 5/7 9: 4 6 6/7 6: 6/7 7: 6/7 : 6/7 9: 4 6 7/7 6: 7/7 7: 7/7 : 7/7 9: 4 6 Model performance typically good: Bias = <2.% R 2 =.9 NSE=.95 Despite limited representation (only by bed heat flux plates) of potential: Lateral groundwater inflows Volumes, residence times & temperatures of hyporheic water Therefore- confident that patterns driven by surface energy exchages

13 1 Results: T w model Water parcels released at m averaged 6 hrs 3 mins to travel 5 m Distance downstream (m) /7 6: 6/7 7: 6/7 : 6/7 9: 1/7 6: 1/7 7: 1/7 : 1/7 9: On days 1 & 3-6 (high energy receipt) parcels released prior to solar noon warmed > 4.2 C 7/7 6: 7/7 7: 7/7 : 7/7 9: between and 5 m Parcels were cooler on arrival at 5 m than parcels concurrently at m. 2/7 6: 2/7 7: 2/7 : 2/7 9: On day 2 (low energy receipt) parcels released prior to solar noon warmed very little between 3/7 6: 3/7 7: and 53/7 m : 3/7 9: Parcels were the same temperature on arrival at 5 m as parcels concurrently at m.

14 Interpretations What generates instantaneous longitudinal cooling gradients? 1. Dramatically reduced energy gains under the riparian canopy in comparison to open landuse, which reduces the rate at which water temperature increases as it flows downstream. 2. While temperature of water advected into the reach increases over morning. Q * Q h Q e Q * Q h Q e Q a Q bhf Q bhf

15 Conclusions Water does not necessarily cool as it flows downstream under seminatural forest canopy in the presence of energy gain conditions, yet longitudinal gradients may still be generated. Energy gains to the water column are reduced dramatically in comparison to open landuse. Longitudinal cooling gradients may be generated when temperature of water advected into a reach increases consistently, while temperature increases are minimal under the forest canopy.

16 Implications Wish to mitigate against high water temperatures? Shade headwaters! Where water hasn t warmed to equilibrium with atmosphere and is thus cooler than majority of locations lower in the basin Can hypothesise that this will also lower temperatures downstream, by reducing longitudinal advection of heat, and so warming

17 Acknowledgements Thank you to: Marine Scotland Freshwater Laboratory staff Jason Leach & Dan Moore NERC Anne Anckorn Nigel Mottram We also have a poster on investigating the effects of shading scenarios on reach-scale water temperature dynamics! Poster 4-2P