AK K64G OLMT Scaling N and P Deposition Demo

ELM requires time series of nitrogen (N) and phosphorus (P) deposition from the atmosphere, in addition to the meteorological time series. Atmospheric deposition of N and P is an important source of these nutrients to vegetation, and in this example, we'll demonstrate how OLMT can be used to look at sensitivity of ELM to changes in these deposition rates.

Example cases doubling N and P deposition

Case 0: Control case to compare against perturbed cases.

First, we run a case with no changes to the N and P deposition data to use as a comparison with the other cases. If you have already run a case for Kougarok using these instructions. If you're not sure if you've run this case, you could always run the following to ensure you have the control case:

docker run -t -i --hostname=docker --user modeluser -v elmdata:/inputdata -v elmoutput:/output serbinsh/ngeearctic_elm_containers:elm_v2-for-ngee_multiarch /scripts/OLMT_docker_example.sh --site_name=kougarok --ad_spinup_years=200 --final_spinup_years=600

Case 1: Double N deposition

In the first case, we double only N deposition, leaving P deposition unchanged. The perturbation starts 5 years into the transient run, e.g., on 1855-01-01:

docker run -t -i --hostname=docker --user modeluser -v elmdata:/inputdata -v elmoutput:/output serbinsh/ngeearctic_elm_containers:elm_v2-for-ngee_multiarch /scripts/OLMT_docker_example.sh --site_name=kougarok --ad_spinup_years=200 --final_spinup_years=600 --scale_ndep=2.0 --startdate_scale_ndep=18550101 --case_prefix=double_Ndep

Case 2: Double P deposition

In the second case, we double only P deposition, leaving N deposition unchanged. The perturbation starts 5 years into the transient run, e.g., on 1855-01-01:

docker run -t -i --hostname=docker --user modeluser -v elmdata:/inputdata -v elmoutput:/output serbinsh/ngeearctic_elm_containers:elm_v2-for-ngee_multiarch /scripts/OLMT_docker_example.sh --site_name=kougarok --ad_spinup_years=200 --final_spinup_years=600 --scale_pdep=2.0 --startdate_scale_pdep=18550101 --case_prefix=double_Pdep

Case 3: Double both N and P deposition

In the final case, we double both N and P deposition, showing the combined effects of doubling both deposition rates. The perturbation starts 5 years into the transient run, e.g., on 1855-01-01:

docker run -t -i --hostname=docker --user modeluser -v elmdata:/inputdata -v elmoutput:/output serbinsh/ngeearctic_elm_containers:elm_v2-for-ngee_multiarch /scripts/OLMT_docker_example.sh --site_name=kougarok --ad_spinup_years=200 --final_spinup_years=600 --scale_pdep=2.0 --startdate_scale_pdep=18550101 --scale_ndep=2.0 --startdate_scale_ndep=18550101 --case_prefix=double_NPdep

Results:

In the elmlab_3.3.2 docker image, there is a script that allows us to compare results from two different simulations: plot_ELM_compare_cases.ipynb. You'll need an elmlab_3.3.2 container running; you may already have one (check in docker desktop). If there isn't one running, you can start one using the instructions here: Analyzing ELM Output in Jupyter Lab.

Comparing case 1 and case 0: impact of doubling nitrogen deposition

Let's take a look at the impact that the impacts doubling nitrogen deposition rates had on ecosystem biogeochemistry. In the second cell of the plot_ELM_compare_cases.ipynb notebook, you'll see there is a place to select the cases we'd like to compare. Let's choose the base case (either OLMT_AK-K64G* or base_AK-K64G if you ran case 0 above) and the future temperature case (double_N_AK-K64G*):

In the next cell, we select the variable and time period we wish to look at across the two cases. Let's start with total ecosystem nitrogen (TOTECOSYSN):

If you continue by running all of the cells, the last cell will produce a plot like this:


This plot shows the total nitrogen content of the ecosystem - the left plot shows values for each of the cases, while the right plot shows the difference. The cases are identical until 1855, but after N deposition is doubled, ecosystem N content increases more rapidly in the experimental case than the control case. The difference is shown in the right panel: from this panel, we can see that the change in total N due to the perturbation is rather modest relative to the total pool size, but that the increase in ecosystem nitrogen across 1850-2015 is ~25% larger when nitrogen deposition is doubled.

Does the change in nitrogen deposition cause changes in ecosystem carbon stocks? We can investigate by going up to the cell where we selected variables, and look at TOTECOSYSC:

Working through each cell, the bottom cell produces a plot like this:

The left plot again shows the total ecosystem carbon pools for both cases - ecosystem carbon increases more rapidly in the experimental case than in the control case. The right plot shows the difference between the two - again, the increase appears fairly modest, but doubling nitrogen deposition causes ~10% greater carbon pool growth relative to the control case.

Take a look at some of the other nitrogen and carbon pools throughout the model - there are a lot of them, as these pools are tracked through soils, leaves, litter, total vegetation mass!

Comparing cases 2 and 0: impact of doubling phosphorus deposition only

What if we double phosphorus deposition rates instead of nitrogen deposition rates? To do this, we have to change the cases we specify as the experimental case. In the top cell, we'll need to specify the 'double_P' case as the experimental case:

As above, let's take a look at some of the total ecosystem pools, starting with total ecosystem phosphorus:

Working through the notebook to the last cell will produce line plots of total ecosystem phosphorus:

In both cases, we see a linear increase in total ecosystem phosphorus over time, with the rate of P increase notably increasing in the experimental case when the phosphorus deposition rate doubles in 1855. However, the monotonic increase in both cases with no variations suggests that this case is not sufficiently spun up, and therefore, the phosphorus pools may not be at equilibrium (see point #3 in the last section of this demo!)

What about the total ecosystem carbon pool? If we go back and select the TOTECOSYSC variable, the last cell in the notebook should produce a plot similar to this:

Comparing cases 3 and 0: doubling both phosphorus and nitrogen deposition

Above we saw that doubling the nitrogen deposition rate caused increases in ecosystem carbon and nitrogen pools, while doubling the phosphorus deposition rate had only a small impact. Which of these effects is stronger, and does ELM predict the soil to become wetter or drier under future conditions? Take a look at plots for soil water content and temperature for the 'future_TP' case relative to case 0.

Your turn:

1. In this case we've increased N and P deposition, but what about if N and P deposition were lower? Try running cases where N and P deposition were halved instead of doubled. Be sure to also change the --case_prefix argument or you will lose your work!

2. Is the sensitivity to changes in P and N deposition the same across sites? Try a similar series of simulations, but by changing the --site_name argument to teller, council, or beo.

3. Try running a 5x longer spinup (note: this will take a while!) - you'll need to change the --ad_spinup_years and --final_spinup_years arguments to the /scripts/OLMT_docker_example.sh script in the docker run calls above.

Return to Home