Calculating water savings from willow removal

CSIRO research has confirmed the long-held suspicion that water can potentially be returned to creeks and streams if riverine infestations of the Salix willow species are removed.

Water with willow trees

Willows overhanging the Ovens Creek in Victoria.

In natural systems in southeast Australia, stream beds are generally unoccupied and willow establishment within streams increases total riparian leaf area which increases total evaporative losses. Within this region, where the CSIRO research was undertaken, willows have been identified as naturalized weeds which invade stream beds.

The data shows that willows account for more lost water than equivalent densities of native species when located in stream beds. Indeed, they remove more water than would occur if there were no vegetation but only open water.

However, while water is potentially returned when riverine willow stands are removed, culling willows from water-limited environments such as river banks is unlikely to generate water savings.

In the past, a key problem for catchment managers has been accurately identifying how much water might be returned if a specific copse of willows was removed. Now, after years of multidisciplinary research, CSIRO has developed a method that helps catchment managers calculate the potential water payback and also improves cost-benefit analyses of willow removal.

Water balance calculations over four years of field research to measure willow water use, indicate that an average net water saving of 3.9 – 5.5 ML a year for each hectare of willow canopy removed could be achieved in cool temperate and semi-arid climates respectively.

fine roots in the water

The prolific roots that form extensive, thick mats which clog creeks and streams in Australia.

The researchers developed monthly pan coefficients from 30 years of meteorological data for 30 key reference sites across Australia, covering broad climatic ranges, for both the Salix babylonica and Salix fragilis willow species. These tables allow managers to estimate evaporative losses from in-stream willow presence and potential water savings from removal based on climatic zone.

To further enhance and improve willow management practices, an economical remote sensing technique was developed using satellite imagery. This technique, relying on spectral discrimination methods, readily discriminates the canopy area of willows from native vegetation. In-stream willows are also identified using delineation of stream bank extent to separate them from bank located willows.

Delineating willow canopy area provides a method to scale-up evapotranspiration and water savings predictions to the catchment perspective, to account for catchment evaporative losses stemming from removal of in-stream willows.

Used in conjunction with the tables in this fact sheet, the relatively low-cost satellite mapping tool enables weed and water managers to make accurate regional estimates of willow evapotranspiration and potential water savings for mapped target areas.

The remote sensing tool also allows catchment managers to monitor willow distribution spatially and temporally, using a method with potential global application.

The results presented provide a guide to potential water savings related to removal of in-stream willows. However, before removing willows, all environmental benefits and disbenifits of removal must be considered, such as the bank stabilisation role that willows provide.

Willows along a river bank

Willows along the River Murray at Blanchetown during a flood.

Behind the science

sapflow sensor and battery installed in a willow tree

Sap flow sensors installed in willows are used to measure how mcuh water willows consume.

Management tools derived from this body of research provide methods to scale willow and native riparian evaporative losses, and associated potential water-savings estimates, from local to regional perspectives. The methodologies presented also have global application.

The first study quantified the various components of willow water use in two semi-arid locations over three years and examined whether water savings could be achieved by willow removal.

In a semi-arid region, evapotranspiration (or water loss) of invasive willows, native riparian vegetation and unshaded open water were compared to estimate water savings from willow removal and replacement with native vegetation (Table A). Three years of willow evapotranspiration data was collected to water saving estimates were accurate. Additional field measurements of willow water use for one year were made at two cool temperate sites for a second willow species (Table B) for comparison with warmer climates.

TABLE A: Measured or estimated water balance components and calculated total evapotranspiration (ET) of Red Gums, Salix babylonica and open water for the 2005/2006 and 2006/2007 measurement period at Jerilderie, and the 2007/2008 measurement period at Yanco. “Savings” refers to potential water savings (ML year-1) per hectare of willow crown projected area, if in-stream willows are removed. 95 % confidence intervals (CI) for each water balance component and total ET are shown, illustrating upper and lower limits of total ET and potential water savings.

Water use site Rainfall(mm) Interception (mm) Evaporation (mm) Transpiration (mm) Total ET (mm) Water savings (ML year-1 ha_1) Water savings CI (mm) Upper CI (mm)/savings (ML year-1 ha_1) Lower CI (mm)/savings (ML year-1 ha_1)
Red Gums 05/06 314 52 ± 5.2 261 ± 14.8 240 ± 78.2 553 ±78 631/— 475/—
S. babylonica
– bank 05/06
314 26 ± 1.8 257 ± 12.8 280 ± 51.1 563 ±51 614/— 512/—
Willows
– creek 05/06
314 26 ± 1.8 751 1633 ± 544.0 2410 +9.4 ±544 2954/14.8 1866/3.9
Open water 05/06 314 1472
(est Morton ET)
1472
S. babylonica 06/07 241 32 ± 2.6 789 ± 71.3 934 ± 328.0 1755 +1.5 ±328 2083/4.8 1427/—
Open water 06/07 241 1604 ± 279.1 1604
S. babylonica
07/08
401 111 ± 9.7 812 ± 64.2 1024 ± 343.0 1947 +5.5 ±343 2290/8.9 1604/2.1
Open water 07/08 401 1396 ± 281.40 1396

TABLE B: Measured water balance components and total evapotranspiration (ET) of Salix fragilis and open water evaporation at Happy Valley (HV) and Tea Garden (TG) and mixed riparian eucalypts at TG in 2008/2009. “Savings” refers to potential water savings (ML year-1) per hectare of willow crown projected area, if in-stream willows are removed. 95 % confidence intervals (CI) for each water balance component and total ET are shown, illustrating upper and lower limits of total ET and potential water savings.

Water use site Rainfall(mm) Interception (mm) Evaporation (mm) Transpiration (mm) Total ET (mm) Water savings (ML year-1 ha_1) Water savings CI (mm) Upper CI (mm)/savings (ML year-1 ha_1) Lower CI (mm)/savings (ML year-1 ha_1)
HV S. fragilis

2008/2009

725 58 (±6) 818 (±141) 340 (±121) 1216 +2.9 ±141 1357/4.3 954/1.5
Open water 2008/2009 926 (±170) 926
TG S. fragilis

2008/2009

474 183 (±54) 702 (±100) 455 (±186) 1340 +4.9 ±186 1528/6.7 1156/3.0

 

Open water 2008/2009 854 (±188) 854
TG riparian 2008/2009 474 192 (±9) 205 (±48) 923 (±498) 1320 +4.6 ±498 1818/9.6 822/—
Open water 2008/2009 854 (±188) 854

The researchers also showed how willow transpiration is related to stomatal conductance and leaf area index.

The collated data was used to calibrate and validate the Penman-Monteith model of evapotranspiration. This model was run for the two willow species using 30 years of climate data at each of 30 locations across seven biogeoclimatic zones of Australia, populating tables of monthly pan coefficients for each location (Table C and Table D).

The pan coefficients provide a straightforward method for water resource managers to estimate long-term monthly and annual willow evapotranspiration based on local measurements of pan evaporation (Table E). Coupled with a simple model of open water evaporation derived from field data, potential water savings for each location can be calculated.

The study found substantial differences in transpiration flux rates between willows with permanent and semi-permanent access to water, with peak transpiration of 15.2 mm/day and 2.3 mm/day respectively. Water balance calculations over the four years indicate that an average net water saving of 3.9 – 5.5 ML a year for each hectare of willows removed could be achieved in cool temperate and semi-arid climates respectively.

In more recent research, a single, very high 2 m-resolution multispectral WorldView-2 satellite image was used to discriminate willow stands from surrounding native riparian vegetation.

Canopy area estimates of in-stream willows in a 25km2 study area, coupled with water savings estimates from willow removal, suggest removal of 10.4 ha of Salix fragilis canopy from within river channels will potentially return 41ML per year to the environment.

A colour photograph of willows along a creek

Aerial image of willows along and in tea Garden Creek near Myrtleford, Victoria. Willows present as a brighter green than native vegetation.

Black and white remote sensing image of a creek with willows mapped

Worldview2 Image with area of in-stream willows mapped in black and white along Tea Garden Creek, near Myrtleford, Victoria.

CASE STUDY

Below we present an example of how to use the tools generated by the research calculate annual willow evapotranspiration (ET) and potential water savings for a location, based on climate using pan coefficients.

  1. To calculate ET for each month (mm month) for location and species decide on a location in Table C, for example Albury and the species is Salix bablyloncia. Next, mulitply each months pan coefficient by the same months pan evaporation for Albury in Table E.
  2. To calculate annual ET (mm year-1) for Albury and S. babylonica, calculate the sum of ET for all months.
  3. To calculate potential water savings (mm year-1) for location and species i.e. Albury and S. bablyloncia subtract annual open water evaporation (last column in Table E) from annual willow ET.

For S.babylonica located in the Albury region, annual ET is 1499 mm year-1, open water evaporation is 981 mm-1 year-1 and potential water saving is 518 mm-1 year-1 per hectare of willow canopy removed.

Similar calculations can be undertaken for S. fragilis using pan coefficients in Table D.

By mapping willow area with the remote sensing method developed, it is then possible to accurately calculate for a region, the potential water savings if willows were to be removed as demonstrated above for 10.4 ha of S. fragilis.

TABLE C: Salix Babylonica pan coefficients for each reference site per month.

Reference sites Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Albany 0.94 0.92 0.96 1.08 1.16 1.18 1.22 1.23 1.24 1.20 1.09 1.01
Ballarat 1.04 1.04 1.12 1.23 1.35 1.48 1.42 1.37 1.32 1.24 1.18 1.12
Beechworth 1.02 1.05 1.13 1.38 1.75 2.10 2.00 1.76 1.54 1.34 1.17 1.10
Bendigo 1.03 1.04 1.10 1.23 1.44 1.63 1.56 1.46 1.35 1.24 1.15 1.10
Canberra 0.97 1.02 1.07 1.25 1.44 1.70 1.64 1.41 1.26 1.17 1.08 1.04
Charlton 0.96 0.97 1.03 1.18 1.36 1.58 1.51 1.41 1.30 1.18 1.08 1.03
Coffs 1.03 1.04 1.08 1.14 1.24 1.30 1.30 1.25 1.16 1.12 1.10 1.08
Colac 0.97 0.97 1.02 1.10 1.17 1.22 1.19 1.17 1.15 1.13 1.09 1.03
Deniliquin 0.92 0.95 1.00 1.13 1.33 1.63 1.53 1.40 1.25 1.15 1.03 0.96
Glen Innes 1.10 1.14 1.17 1.31 1.52 1.76 1.63 1.41 1.23 1.17 1.14 1.13
Hamilton 0.96 0.94 1.00 1.15 1.32 1.44 1.39 1.32 1.26 1.19 1.12 1.05
Horsham 0.92 0.93 0.98 1.13 1.29 1.46 1.39 1.32 1.20 1.13 1.05 1.00
Lake Eildon 1.20 1.21 1.30 1.54 1.78 1.93 1.83 1.74 1.60 1.49 1.38 1.30
Laverton 0.99 0.97 1.04 1.12 1.16 1.31 1.25 1.23 1.21 1.15 1.11 1.06
Mildura 0.86 0.87 0.92 1.03 1.18 1.33 1.30 1.17 1.05 0.99 0.93 0.91
Moorabbin 1.10 1.09 1.15 1.26 1.34 1.39 1.34 1.34 1.32 1.28 1.23 1.18
Moree 0.90 0.91 0.95 1.03 1.17 1.30 1.29 1.18 1.07 1.00 0.95 0.91
Mount Gambier 0.97 0.95 0.98 1.06 1.20 1.25 1.23 1.18 1.17 1.15 1.09 1.05
Narrandera 0.90 0.94 1.00 1.16 1.38 1.60 1.61 1.46 1.29 1.17 1.03 0.97
Orbost 1.20 1.21 1.29 1.39 1.49 1.52 1.47 1.34 1.30 1.31 1.28 1.25
Ouyen 0.90 0.92 0.97 1.10 1.25 1.43 1.40 1.26 1.13 1.05 0.99 0.95
Perth 0.91 0.91 0.94 1.07 1.20 1.29 1.33 1.35 1.28 1.17 1.04 0.97
Portland 0.89 0.88 0.93 1.08 1.23 1.32 1.31 1.27 1.25 1.19 1.09 0.99
Rutherglen 0.95 0.98 1.04 1.23 1.49 1.74 1.68 1.53 1.36 1.21 1.08 1.01
Sale 0.98 0.97 1.03 1.16 1.27 1.35 1.27 1.24 1.17 1.15 1.11 1.02
Stawell 0.96 0.96 1.03 1.20 1.38 1.60 1.51 1.48 1.38 1.23 1.13 1.05
Strathgordon 1.25 1.26 1.31 1.46 1.51 1.55 1.59 1.55 1.48 1.41 1.42 1.29
Swan Hill 0.95 0.97 1.02 1.15 1.33 1.54 1.48 1.35 1.21 1.13 1.03 1.01
Tatura 0.97 0.99 1.06 1.20 1.39 1.62 1.51 1.40 1.29 1.19 1.08 1.01
Wagga 0.91 0.95 1.00 1.18 1.41 1.66 1.65 1.52 1.34 1.17 1.03 0.97

TABLE D: Salix Fragilis pan coefficients for each reference site per month.

Reference sites Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Albany 0.79 0.76 0.78 0.83 0.83 0.76 0.73 0.62 0.77 0.91 0.87 0.83
Ballarat 0.87 0.86 0.89 0.93 0.92 0.88 0.78 0.64 0.80 0.92 0.94 0.91
Beechworth 0.86 0.87 0.91 1.06 1.22 1.29 1.12 0.84 0.94 1.00 0.94 0.90
Bendigo 0.86 0.86 0.88 0.94 1.00 0.99 0.87 0.70 0.83 0.93 0.92 0.89
Canberra 0.83 0.85 0.86 0.95 1.01 1.06 0.94 0.68 0.77 0.87 0.87 0.86
Charlton 0.80 0.80 0.83 0.90 0.94 0.97 0.85 0.68 0.80 0.88 0.86 0.83
Coffs 0.89 0.89 0.90 0.90 0.92 0.89 0.82 0.66 0.75 0.86 0.90 0.91
Colac 0.81 0.80 0.81 0.83 0.79 0.73 0.65 0.55 0.69 0.84 0.87 0.84
Deniliquin 0.77 0.78 0.80 0.87 0.94 1.03 0.87 0.69 0.78 0.86 0.82 0.78
Glen Innes 0.94 0.95 0.95 1.01 1.10 1.16 0.98 0.71 0.77 0.88 0.93 0.94
Hamilton 0.80 0.77 0.80 0.87 0.89 0.86 0.76 0.62 0.76 0.88 0.89 0.86
Horsham 0.77 0.76 0.79 0.85 0.89 0.89 0.77 0.63 0.74 0.84 0.84 0.81
Lake Eildon 1.01 1.00 1.05 1.17 1.22 1.15 1.00 0.81 0.97 1.11 1.11 1.07
Laverton 0.83 0.80 0.83 0.84 0.79 0.79 0.69 0.58 0.73 0.85 0.88 0.86
Mildura 0.70 0.70 0.73 0.79 0.84 0.85 0.77 0.59 0.65 0.74 0.73 0.73
Moorabbin 0.93 0.90 0.92 0.95 0.91 0.84 0.74 0.64 0.80 0.95 0.98 0.96
Moree 0.76 0.77 0.77 0.81 0.87 0.88 0.80 0.62 0.68 0.76 0.76 0.75
Mount Gambier 0.82 0.78 0.78 0.80 0.81 0.76 0.69 0.57 0.71 0.86 0.87 0.85
Narrandera 0.75 0.77 0.80 0.90 0.99 1.02 0.94 0.72 0.81 0.88 0.82 0.79
Orbost 1.01 1.01 1.05 1.07 1.03 0.91 0.79 0.66 0.81 0.98 1.03 1.03
Ouyen 0.74 0.74 0.77 0.84 0.89 0.91 0.82 0.63 0.71 0.79 0.78 0.77
Perth 0.76 0.74 0.76 0.83 0.88 0.87 0.82 0.71 0.81 0.89 0.84 0.80
Portland 0.75 0.72 0.74 0.80 0.83 0.80 0.73 0.61 0.76 0.89 0.87 0.81
Rutherglen 0.75 0.77 0.79 0.90 1.00 1.03 0.91 0.71 0.82 0.88 0.82 0.77
Sale 0.83 0.80 0.83 0.88 0.86 0.82 0.71 0.59 0.71 0.86 0.88 0.83
Stawell 0.80 0.79 0.83 0.91 0.95 0.96 0.83 0.71 0.84 0.92 0.90 0.85
Strathgordon 1.03 1.02 1.01 1.04 0.95 0.84 0.79 0.65 0.85 1.02 1.11 1.03
Swan Hill 0.79 0.79 0.82 0.89 0.93 0.97 0.86 0.67 0.75 0.84 0.82 0.81
Tatura 0.82 0.82 0.86 0.91 0.96 0.99 0.84 0.66 0.79 0.89 0.87 0.83
Wagga 0.76 0.78 0.81 0.91 1.00 1.05 0.95 0.74 0.83 0.88 0.82 0.79

TABLE E: Pan evaporation (mm day-1) for each reference site, averaged for the period 1980-2010. Annual modelled open water evaporation per reference site (mm year-1) is given in the last column.

Reference sites Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Open E (mm year-1)
Albany 6.69 6.15 4.91 3.24 2.24 1.79 1.79 2.27 2.88 3.69 4.91 6.11 981
Ballarat 6.30 5.80 4.21 2.62 1.51 0.98 1.06 1.63 2.41 3.49 4.65 5.51 843
Beechworth 7.44 6.38 4.74 2.62 1.34 0.80 0.83 1.34 2.24 3.56 5.19 6.44 901
Bendigo 7.60 6.87 5.08 3.07 1.68 1.07 1.14 1.78 2.71 4.11 5.71 6.78 999
Canberra 7.21 5.97 4.71 2.95 1.80 1.19 1.30 2.08 3.17 4.38 5.62 6.57 987
Charlton 8.35 7.51 5.54 3.34 1.87 1.18 1.27 1.95 2.96 4.48 6.26 7.42 1095
Coffs 5.77 5.12 4.40 3.46 2.53 2.15 2.37 3.20 4.37 4.99 5.34 5.63 1038
Colac 6.23 5.71 4.29 2.76 1.75 1.25 1.36 1.98 2.73 3.70 4.70 5.53 882
Deniliquin 8.95 8.03 6.01 3.56 1.98 1.26 1.35 2.02 3.19 4.75 6.94 8.42 1186
Glen Innes 5.35 4.60 4.02 2.81 1.83 1.29 1.47 2.28 3.58 4.37 4.84 5.23 877
Hamilton 6.68 6.25 4.61 2.76 1.58 1.08 1.19 1.77 2.51 3.54 4.71 5.70 890
Horsham 8.11 7.36 5.41 3.26 1.85 1.21 1.32 1.93 2.90 4.25 5.93 7.10 1063
Lake Eildon 5.77 5.24 3.81 2.23 1.24 0.82 0.89 1.36 2.13 3.14 4.22 5.04 754
Laverton 6.63 6.09 4.56 3.00 1.96 1.37 1.53 2.08 2.89 3.97 5.02 6.02 948
Mildura 10.38 9.28 7.05 4.41 2.57 1.73 1.84 2.89 4.46 6.35 8.34 9.49 1445
Moorabbin 5.87 5.41 4.05 2.65 1.71 1.28 1.40 2.00 2.75 3.70 4.59 5.28 855
Moree 9.12 8.04 6.93 4.85 3.17 2.25 2.32 3.38 5.06 6.70 8.16 8.97 1450
Mount Gambier 6.55 6.06 4.53 2.85 1.71 1.25 1.36 1.95 2.68 3.64 4.78 5.68 904
Narrandera 9.58 8.14 6.23 3.71 2.07 1.27 1.28 2.02 3.22 4.95 7.13 8.56 1222
Orbost 5.65 4.93 3.82 2.59 1.71 1.34 1.45 2.05 2.86 3.72 4.50 5.09 835
Ouyen 9.56 8.54 6.43 3.98 2.33 1.53 1.63 2.54 3.85 5.63 7.55 8.70 1308
Perth 9.44 8.60 6.90 4.38 2.83 2.04 1.99 2.45 3.34 4.90 6.86 8.35 1304
Portland 6.35 5.95 4.44 2.77 1.68 1.24 1.34 1.91 2.59 3.51 4.56 5.47 878
Rutherglen 8.51 7.26 5.44 3.12 1.68 1.05 1.09 1.68 2.67 4.11 6.02 7.48 1052
Sale 6.16 5.59 4.15 2.68 1.75 1.35 1.52 2.02 2.93 3.76 4.72 5.72 890
Stawell 7.53 6.85 5.02 3.02 1.68 1.07 1.17 1.70 2.56 3.87 5.38 6.52 974
Strathgordon 4.02 3.53 2.53 1.49 0.96 0.67 0.70 1.07 1.67 2.48 3.19 3.81 549
Swan Hill 8.98 7.99 6.03 3.69 2.11 1.36 1.46 2.28 3.48 5.16 7.05 8.12 1212
Tatura 7.75 7.00 5.10 3.01 1.66 1.09 1.17 1.76 2.76 4.08 5.85 7.20 1017
Wagga 9.32 7.88 6.01 3.52 1.94 1.17 1.19 1.81 2.86 4.52 6.71 8.21 1158

Contact Person

Additional Information