*4.3. Experiment 3*

The input of nitrogen to surface waters is increasing throughout the baseline simulation starting at an initial rate of 3.1 trillion moles or 4.3 Mt per year to a rate of 7.6 trillion moles or 10.5 Mt per year (Figure 8). Input of phosphorus to surface waters on the other hand, increases from 451 billion moles or 13.5 Mt per year to a peak value of 681 billion moles or 20.4 Mt per year in the year 2025. After this point phosphorus input decreases significantly, down to 126 billion moles per year.

**Figure 8.** Total nitrogen and phosphorus input to surface water under the ANEMI3 baseline scenario. Left axis represents number of moles of nitrogen and phosphorus inputs to surface water per year.

The explanation for the difference in the pattern of nitrogen and phosphorus inputs lies in their respective amounts in different sources. For nitrogen, on a global scale, agriculture is the main anthropogenic source of nutrients to surface waters, while domestic and industrial wastewaters are the main source of phosphorus. Phosphorus input decreases after the year 2025 due to increasing levels of wastewater treatment on a global scale, which reduces the input significantly. The levels of treated and untreated wastewater are shown in Figure 9. Initially, the amount of untreated wastewater is greater than treated on a global scale in 1980. Under the ANEMI3 baseline scenario, wastewater treatment increases from the initial rate of 160 km<sup>3</sup>/year and surpasses that of the untreated percentages in 2010. After this point, treatment rate increases further to approximately 550 km<sup>3</sup>/year.

**Figure 9.** Treated and untreated wastewater inputs to the nutrient cycles over time.

Nutrient inputs act as an additional rate that affects the surface water stock in the nutrient cycle model. Combining this with the stock of surface water in the hydrologic cycle model allows for the concentrations of nutrients in surface water on a global scale to be examined, as shown in Figure 10. The concentration considers changes in hydrologic cycle. The patterns are almost the same because the global amount of streamflow does not change very much due to climate change increase and surface water consumption having a balancing effect in the ANEMI3 baseline scenario.

**Figure 10.** Surface water nutrient concentrations of nitrogen and phosphorus.

Nutrient concentrations are higher when constant wastewater treatment is implemented, rather than exogenous increase in the ANEMI3 baseline scenario. Nutrient concentrations are used as an indicator for water quality in the production of surface water supplies, whereby higher concentrations act as a multiplier to the surface water production costs. The effect of constant wastewater treatment on water supply development is shown in Figure 11. Under this scenario, the establishment of surface water supplies is only slightly affected by the change in surface water quality on a global scale (Figure 11a). Under the ANEMI3 baseline parameterization scheme, water quality does not appear to play a significant role in the establishment of surface water supplies, even if wastewater treatment levels are held at constant 1980 values for the entire simulation. Both wastewater reuse and desalination supplies show major increases from 1980 to the year 2100. Wastewater reuse increases from 10 to 280 km<sup>3</sup>/year, while desalination increases from 10 to 75 m<sup>3</sup>/year, although the absolute numbers are small in comparison to conventional water supplies. With reduced wastewater treatment rates there is a major difference in the level of wastewater reuse, as there is less available wastewater resource to be used (Figure 11b). Due to scarce wastewater for reuse there is a drop from 274 km<sup>3</sup>/year to 143 km<sup>3</sup>/year by the year 2100.

**Figure 11.** Development of water supplies under the baseline and constant wastewater treatment scenarios for (**a**) conventional water supplies and (**b**) alternative water supplies.
