SAC Soil Moisture Workshop. 25 th January Contact:

Transcription

1 COSMOS-UK: A New Field-scale National Soil Moisture Measurement Network Jonathan Evans, Lucy Ball, James Blake, Olivia Hitt, Dan Rylett, George Wright, Hollie Cooper, Helen Vincent, Sarah Bagnoli, Pete Scarlett, Ross Morrison, Matt Fry, David Boorman, Andrew Singer and Richie Ellis. Centre for Hydrology & Ecology, Wallingford, Oxon., UK SAC Soil Moisture Workshop 25 th January 2017 Contact: CosmosUK@ceh.ac.uk

2 Introduction Background & Rationale COSMOS measurement principle Station design measurements & sensors Data Results and future work

3 Scientific Rationale

4 Why is Soil Moisture Important? Because: 1) Soil Moisture (SM) controls the land surface energy balance: Solar radiation (sunlight) warms the earth s surface warming the ground (soil heat flux) - Warming the air (sensible heat flux) - And evaporating moisture from soil and plants (evapotranspiration) - Soil thermal properties are controlled by SM - Reflected and emitted radiation depends on soil moisture

5 Why is Soil Moisture Important? Because: 2) Soil Moisture (SM) controls hydrological processes - Runoff - Infiltration - Recharge of groundwater - Evaporation - Availability of water to plants/crops - Available water resource Diagram from

6 Cosmic Rays and Soil Moisture (1) A cascade of high-energy (~1 GeV) secondary neutrons are produced by primary cosmic ray protons entering the earth s magnetosphere (2) High-energy secondary neutrons collide with nuclei in the atmosphere evaporating fast neutrons (~1MeV) these are scattered in the air and ground, loosing energy to become thermal (0.025 ev) or epithermal neutrons (> 0.5 ev) From Zreda et. al. Hydrol. Earth Syst. Sci., 16, , doi: /hess

7 The COsmic-ray Soil Moisture Observing System Naturally occurring high energy neutrons generated by cosmic rays, are at equilibrium above the surface: the balance of supply of fast neutrons and their moderation (thermalisation) by surrounding nuclei. BF 3 or He 3 gas discharge tube a.k.a. cosmic ray soil moisture probe) detects reduction in fast neutrons as soil moisture increases (more H). atoms). Courtesy of: Marek Zreda, Jim Shuttleworth, Xubin Zeng, Chris Zweck and Ty Ferre Cosmic Ray Soil Moisture Probe (Hydroinnova CRS-2000/B) at Hollin Hill, N. Yorks.

8 COSMOS Measurement area (modelled) COSMOS integrates soil moisture over a 200 m radius 86% of neutrons detected are from within a radius of 200 m Footprint is mostly independent of soil moisture Footprint increases with altitude (decreasing pressure) By Marek Zreda

9 Measurement depth (modelled) COSMOS integrates soil moisture from the surface to a depth of cm depending on the soil moisture content 86% of neutrons from within 70 cm soil depth for very dry conditions In very wet soils only neutrons from the top 12 cm reach detector neutron source soil depths are independent of altitude (pressure) By Marek Zreda

10 Corrections to Neutron Counts Raw neutron counts (N raw ) and the correction factors applied (F p, F Q, F C, black lines) to obtain corrected counts (N corr ) for CHIMN. Relevant variables (p, Q, C) are also shown in each case (green lines, right-hand axes). The temporal resolution of the data is 60 min.

11 Sensors Goal is to measure water and energy balances research grade weather station, 4-component radiometer, heat flux plates and soil temperature profile Point soil moisture sensors (x2) at 10 cm, and soil moisture profile for sense check. High accuracy, low maintenance weighing rain gauge. All sensors logged on a Campbell CR3000 with GPRS telemetry.

12

13 Phase 1 Installation Works Acclima TDT & Soil heat flux plate Gill Metpak OTT Pluvio II install at SHEEP Hukseflux STP01 soil temperature profile IMKO installation, CHIMN Chimney Meadows, Bampton, Oxfordshire (CHIMN)

14 Completed Phase 1 Sites Waddesdon Manor Farm, Buckinghamshire (WADDN) Note: only moderated CRS-1000 tube installed Sheepdrove Organic Farm, Lambourn, West Berkshire (SHEEP)

15

16 WADDESDON Current network UK wide coverage Varying spatial density Range of climate, land cover, soils, geology, topography Simple grassland Build on existing research / networks Pragmatic approach Installed Total 41

17 Data Flows Automated telemetry of data in real-time Automated calibration & quality control Data continuously archived to CEH Oracle database Data requests currently served manually

18 Web resources

19 Other soil moisture sensors (1) Daily data from Rothamsted

20 COSMOS Data Daily averages (lines) and 6-h running means (shading) of COSMOS soil moisture content and modelled effective measurement depth

21 COSMOS & ASCAT SM Indices Soil moisture indices from COSMOS and ASCAT data for Chimney Meadows and Sheepdrove Organic Farm (August 2011-December 2013). The resolution of the data is daily. Dashed lines are 1:1; solid lines are linear regressions through the data.

22 CRS SM compared with JULES VWC from CRS probes compared to 10 cm SM from the JULES model for Chimney Meadows and Sheepdrove Organic Farm J. G. Evans et. al., Soil water content in southern England derived from a cosmic-ray soil moisture observing system COSMOS- UK Hydrological Processes

23 Phenocam Example Time Lapse Movie Moor House Useful for interpreting physical measurements Gives information on land cover, crop growth, grazing etc. Possible uses for biodiversity?

24 Thank you. Questions..?

25 Data Assimilation: COSMIC The variable measurement depth of the Cosmic-ray Soil Moisture Sensor compared to the constant soil moisture depth layers in land surface models (LSM) could be a disadvantage. But by using the predicted soil moisture from the LSM to predict the expected neutron count circumvents this issue. The COsmic-ray Soil Moisture Interaction Code (COSMIC 1 ) provides an analytical model to predict neutron counts from the LSM, allowing data assimilation of measured counts. 1.Shuttleworth et.al. 2013, HESS

26 WHY SOIL MOISTURE IS IMPORTANT? Outline - MOTIVATIONS - THE JULES MODEL - a land-surface model (Thanks to Eleanor Blyth and Alberto Martinez-de la Torre) - THE GRID-TO-GRID MODEL - a hydrological model (Thanks to Vicky Bell)

27 WHY SOIL MOISTURE IS IMPORTANT? Soil moisture effects many human activities Floods Agriculture Land-use Soil availability

28 WHY SOIL MOISTURE MODELLING? Traditional field measurement of soil moisture is time consuming and expensive Soil moisture is not routinely monitored over the long term like precipitation and discharge The remotely sensed soil moisture data has its own uncertainties; e.g. mismatch in scale between the in situ measurements; surface estimates only. vs Increasing computational speeds allow for more accurate numerical modelling techniques Models can tell what will be: evaluating effects of land-use, climate changes

29 THE LAND-SURFACE MODEL JULES JULES is a community model of land surface, vegetation and soil processes led by CEH and the UK Met Office Applications: ~200 users across NERC, Met Office, universities, overseas, >40 NERC proposals since 2008 with JULES activity Used by the Met Office weather and climate models (IPCC) Will be used in NERC/Met Office UK Earth System Model (UKESM1)

30 THE LAND-SURFACE MODEL JULES

31 THE LAND-SURFACE MODEL JULES Energy Budget Energy flows between the surface and the atmosphere: radiation, heat and latent heat Carbon Cycle Carbon assimilated from the air, converted Into leaf, roots and stem carbon. Photosynthesis Carbon and water flows through the stomata between the surface and the atmosphere: photosynthesis, respiration and transpiration Water Budget Water flows between the atmosphere and the surface The snow pack can have up to 3 layers Water and heat flows: ice formation effect CO2 and CH4 release from soil modified from: JULES: Best et al, 2011, Clark et al, 2011, GMD.

32 THE LAND-SURFACE MODEL JULES Energy Budget Energy flows between the surface and the atmosphere: radiation, heat and latent heat Plants grows and compete with each other for light, changing the land cover Carbon Cycle Carbon assimilated from the air, converted Into leaf, roots and stem carbon. Photosynthesis Carbon and water flows through the stomata between the surface and the atmosphere: photosynthesis, respiration and transpiration Water Budget Water flows between the atmosphere and the surface The snow pack can have up to 3 layers Water and heat flows: ice formation effect CO2 and CH4 release from soil modified from: JULES: Best et al, 2011, Clark et al, 2011, GMD.

33 THE LAND-SURFACE MODEL JULES Model INPUT Incoming solar radiation Relative humidity Atmospheric pressure Air temperature Precipitation Wind speed Soil moisture and runoff component Model soil moisture OUTPUT Averaged soil moisture typically at 4 depths 1. Incoming moisture is split into runoff and water absorbed. Runoff is diverted in rivers. 2. There is a constant redistribution of water within the soil column as it tries to reach a state of equilibrium. This is determined using the Darcy s law: q = K Ψ z At the bottom of the soil layers (3m), water is taken out at a rate assuming only gravitational effects free drainage. 4. This drainage joins the surface runoff in rivers. 10 cm 35 cm 100 cm 300 cm

34 THE LAND-SURFACE MODEL JULES Soil moisture content for all UK at 1km resolution: mean for the period Standard deviation of the soil moisture content anomalies at global scale at ~50km resolution for the cm depth cm depth Available at Plot scale soil moisture content evolution in time for different layers. CHESS (1km resolution) driving data available thought the CEH webpage

35 The Grid-to-Grid hydrological model Energy Budget Energy flows between the surface and the atmosphere: radiation, heat and latent heat Photosynthesis Carbon and water flows through the stomata between the surface and the atmosphere: photosynthesis, respiration and transpiration Water Budget Water flows between the surface and the atmosphere: precipitation Carbon Cycle Carbon assimilated from the air, converted Into leaf, roots and stem carbon. CO2 and CH4 release from soil Dry snow Wet snow Unsaturated zone Saturated zone The snow pack can have up to 2 layers Surface and sub-surface runoff routed laterally to estimate downstream rive flows

36 The Grid-to-Grid hydrological model A spatially distributed grid-based model Driven by gridded rainfall and potential evaporation data Estimates naturalised river flows on a 1km grid at 15 min t-step Gridded spatial datasets of landscape properties (soil, land-cover, topography, geology) reduces the need for calibrating model parameters Data available on request: Soil column depth, L River s 0 Saturationexcess surface runoff, q Lateral drainage, Q D Percolation, Q P Groundwater flow (subsurface runoff), Q G daily/monthly/annual 1km grids of depth averaged soil moisture monthly grids should be freely available soon. Subsurface flow Bell et al. (2009). Journal of Hydrology, 377 (3-4),

37 The Grid-to-Grid applications Flood forecasting - used countrywide for operational flood forecasting and warning for the EA and SEPA Models national river flows 24/7, forecasts out to 5 days Surface water flooding Estimating projected future change in UK river flows (collaboration with Met Office Hadley Centre) Seasonal forecasting of river flows and subsurface water storage: Sensitivity to rainfall map used in EA/FFC s Flood Outlook (used by operational flood managers) Flood forecasting with G2G Soil moisture (%) Soil moisture (%)

38 Applications: Sensitivity to rainfall map Bell VA, Davies HN, Kay AL, Marsh TJ, Brookshaw A, Jenkins A Developing a large-scale water-balance approach to seasonal forecasting: application to the 2012 drought in Britain. Hydrological Processes, 27(20), Soil moisture This map uses the G2G estimates of subsurface storage to highlight areas where water storage, s, is below or above the monthly mean (s mean ), and by how much (mm) This map highlights areas where the ground is WET relative to the long term now used by EA/FFC in their fortnightly Flood Outlook Collaboration with EA and Met Office

39 Applications: seasonal hydrological forecasts The most recent end of month G2G sub-surface storage estimate is used as the initial condition for a water-balance forecast of the next 1- and 3- months sub-surface storage using Met Office GloSea5 rainfall forecast ensemble members and climatological PE as input (Bell et al. 2013). G2G run Observed rain+ PE Present day WB model run for 1-3 months ahead Ensemble forecast monthly rain + mean PE Corresponding ensembles of regional river flow estimates for the next 1- and 3-months ahead can be estimated using the water balance (WB) hydrological model using historical and spatial information from the G2G. Bell, V. A., Davies, H. N., Kay, A. L., Brookshaw, A. & A.A, Scaife. A national-scale seasonal hydrological forecast system: Development and evaluation over Britain, HESS (2017, in preparation)

40 To summarise. Land Surface Models (e.g. JULES) Self-consistent representation of energy and evaporation Models many processes (Temperature, CO 2, snow etc.) Complex formulation, slow, can lack accuracy Limited assessment against observations P Direct runoff Surface storage S 2 E q Surface s runoff q Probabilitydistributed soil moisture storage S 1 Recharg e Groundwat er storage S 3 q b Baseflow Hydrological Models (e.g. Grid-to-Grid, PDM) Required to be fast and accurate Realistic representation of river flow processes Good for modelling floods & droughts in river basins Assessed against observations Often used for research into large scale feedbacks Often used for civil engineering applications e.g. flood forecasting

41 To summarize. hydrology Land Surface Models (e.g. JULES) Self-consistent representation of energy and evaporation Models many processes (Temperature, CO 2, snow etc.) Complex formulation, slow, can lack accuracy Limited assessment against observations P Direct runoff Surface storage S 2 E q Surface s runoff q Probabilitydistributed soil moisture storage S 1 Recharg e Groundwat er storage S 3 q b Baseflow Hydrological Models (e.g. Grid-to-Grid, PDM) Required to be fast and accurate Realistic representation of river flow processes Good for modelling floods & droughts in river basins Assessed against observations evaporation Often used for research into large scale feedbacks Often used for civil engineering applications e.g. flood forecasting