Water shortages in California are a growing concern amidst ongoing drought, earlier spring snowmelt, projected future climate warming, and currently mandated water-use restrictions. Increases in population and land use in coming decades will place additional pressure on already limited available water supplies.
We used a state-and-transition simulation model to project future changes in developed (municipal and industrial) and agricultural land use to estimate associated water use demand from 2012-2062. Under current efficiency rates, total water use was projected to increase 1.8 billion cubic meters (+4.1%) driven primarily by urbanization and shifts to more water intensive crops. Only if currently mandated 25% reductions in municipal water use are continuously implemented would water demand in 2062 balance to water use levels in 2012.
This is the first modeling effort of its kind to examine regional land-use related water demand incorporating historical trends of both developed and agricultural land uses.
Project Description
The aim of this research was to quantify future land-use related water demand in California under a “business-as-usual” (BAU) scenario. The BAU scenario assumes LULC dynamics from the historical period persist into the projected period, including historical rates of land conservation. We compiled historical land use and land cover (LULC) change data as well as water use information from both remote sensing and tabular datasets for use in a state-and-transition simulation model (STSM).
Changes in LULC were modeled annually at 1 km by 1 km spatial resolution for the baseline period (1992-2012) using the historical data. For the projection period (2012-2062), the model randomly sampled from the historical distribution of LULC change while tracking water use for developed (municipal and industrial) and agricultural (annual and perennial cropland) land uses.
Methods
We used the LUCAS STSM to project land-use change over a 70 year period (1992-2062) across 40 Monte Carlo simulations to estimate associated water use demand in Mediterranean California. Projections of land use were developed under BAU conditions where future changes were based on recent historical rates of land-use change and land protection.
The LUCAS model is a form of non-stationary, Markov Chain model, where the landscape is divided into a set of simulation cells with each cell assigned a discrete state (i.e. LULC class). Transitions targets were developed using a time-series of historical data describing the rate of change between land-use and land-cover classes and were used within the model to move cells between states over time. For a more thorough description of the STSM framework see Sleeter, Liu, Daniel et al.
The resulting model output included annual LULC and water use projections to the year 2062 across 40 Monte Carlo simulations. Examination of land-use related estimates of future water demand in California are needed to develop effective water resource management plans, given highly variable inter-annual supplies and future climate uncertainty.
Findings and Results
Land-use
- Between 2012 and 2062, developed land cover was projected to increase 62.9% from an average 12,978 km2 to an average 21,141 km2.
- Annual cropland was projected to decline an average 30.3% (8,822 km2).
- Conversion of annual cropland into perennial cropland and encroachment of perennial crops into rangeland resulted in perennial cropland increasing 39.1% (5,192 km2).
- Overall, anthropogenic land uses increased over 8.2% (4,533 km2) from 2012 levels at the expense of rangelands.
- Total cropland area declined 8.6%.
- Rangelands continued to decline (-7.3%) despite the addition of 3,211 km2 of protected rangeland in the BAU scenario.
a) Projected land-use and land-cover (LULC) change for the historical period (1992 – 2012) and the projected period (2012 – 2062) in California’s Central Valley and Oak Woodlands regions under a business-as-usual (BAU) scenario. The 2012 and 2062 LULC maps represent one out of 40 possible Monte Carlo iterations modeled for each time step. See table S1 for a full explanation of the LULC classification scheme. (b) Trends in mean LULC change over the historical and projected period by LULC class (Figure 2 a,b in Wilson et al. 2016).
Water-use
- By 2062, water use was projected to increase by 1.8 billion cubic meters (Bm3; +4.1%) over current use estimates.
- Developed sector water use demand was projected to increase 4.6 Bm3 (+59.1%).
- Water use in annual cropland was projected to decline nearly 30.2% or an estimated 7.3 Bm3 (range of -6.8 to -7.9 Bm3)
- Perennial cropland water use was projected to increase by 4.5 Bm3 (range of 3.9 to 5.1 Bm3) or 37.5%.
- Total cropland water use was projected to decline 2.8 Bm3 from an average 36.2 Bm3 in 2012 to 33.4 Bm3 in 2062 representing a 7.8% decrease in agriculture water use.
For more information visit the Land Use and Climate Change website.
Water shortages in California are a growing concern amidst ongoing drought, earlier spring snowmelt, projected future climate warming, and currently mandated water-use restrictions. Increases in population and land use in coming decades will place additional pressure on already limited available water supplies.
We used a state-and-transition simulation model to project future changes in developed (municipal and industrial) and agricultural land use to estimate associated water use demand from 2012-2062. Under current efficiency rates, total water use was projected to increase 1.8 billion cubic meters (+4.1%) driven primarily by urbanization and shifts to more water intensive crops. Only if currently mandated 25% reductions in municipal water use are continuously implemented would water demand in 2062 balance to water use levels in 2012.
This is the first modeling effort of its kind to examine regional land-use related water demand incorporating historical trends of both developed and agricultural land uses.
Project Description
The aim of this research was to quantify future land-use related water demand in California under a “business-as-usual” (BAU) scenario. The BAU scenario assumes LULC dynamics from the historical period persist into the projected period, including historical rates of land conservation. We compiled historical land use and land cover (LULC) change data as well as water use information from both remote sensing and tabular datasets for use in a state-and-transition simulation model (STSM).
Changes in LULC were modeled annually at 1 km by 1 km spatial resolution for the baseline period (1992-2012) using the historical data. For the projection period (2012-2062), the model randomly sampled from the historical distribution of LULC change while tracking water use for developed (municipal and industrial) and agricultural (annual and perennial cropland) land uses.
Methods
We used the LUCAS STSM to project land-use change over a 70 year period (1992-2062) across 40 Monte Carlo simulations to estimate associated water use demand in Mediterranean California. Projections of land use were developed under BAU conditions where future changes were based on recent historical rates of land-use change and land protection.
The LUCAS model is a form of non-stationary, Markov Chain model, where the landscape is divided into a set of simulation cells with each cell assigned a discrete state (i.e. LULC class). Transitions targets were developed using a time-series of historical data describing the rate of change between land-use and land-cover classes and were used within the model to move cells between states over time. For a more thorough description of the STSM framework see Sleeter, Liu, Daniel et al.
The resulting model output included annual LULC and water use projections to the year 2062 across 40 Monte Carlo simulations. Examination of land-use related estimates of future water demand in California are needed to develop effective water resource management plans, given highly variable inter-annual supplies and future climate uncertainty.
Findings and Results
Land-use
- Between 2012 and 2062, developed land cover was projected to increase 62.9% from an average 12,978 km2 to an average 21,141 km2.
- Annual cropland was projected to decline an average 30.3% (8,822 km2).
- Conversion of annual cropland into perennial cropland and encroachment of perennial crops into rangeland resulted in perennial cropland increasing 39.1% (5,192 km2).
- Overall, anthropogenic land uses increased over 8.2% (4,533 km2) from 2012 levels at the expense of rangelands.
- Total cropland area declined 8.6%.
- Rangelands continued to decline (-7.3%) despite the addition of 3,211 km2 of protected rangeland in the BAU scenario.
a) Projected land-use and land-cover (LULC) change for the historical period (1992 – 2012) and the projected period (2012 – 2062) in California’s Central Valley and Oak Woodlands regions under a business-as-usual (BAU) scenario. The 2012 and 2062 LULC maps represent one out of 40 possible Monte Carlo iterations modeled for each time step. See table S1 for a full explanation of the LULC classification scheme. (b) Trends in mean LULC change over the historical and projected period by LULC class (Figure 2 a,b in Wilson et al. 2016).
Water-use
- By 2062, water use was projected to increase by 1.8 billion cubic meters (Bm3; +4.1%) over current use estimates.
- Developed sector water use demand was projected to increase 4.6 Bm3 (+59.1%).
- Water use in annual cropland was projected to decline nearly 30.2% or an estimated 7.3 Bm3 (range of -6.8 to -7.9 Bm3)
- Perennial cropland water use was projected to increase by 4.5 Bm3 (range of 3.9 to 5.1 Bm3) or 37.5%.
- Total cropland water use was projected to decline 2.8 Bm3 from an average 36.2 Bm3 in 2012 to 33.4 Bm3 in 2062 representing a 7.8% decrease in agriculture water use.
For more information visit the Land Use and Climate Change website.