Environmental Change Model
What is the ECM?
The Environmental Change Model (ECM) estimates snow/ice mass balance and potential biomes across the globe for climate boundary conditions ranging from Last Glacial Maximum (LGM; ~ 20,0000 years ago) to 2100 CE. Solutions are calculated from gridded inputs of monthly temperature and precipitation using a degree day solver and biome rubric. Boundary conditions for a given experiment are derived by blending reanalysis (modern climate) and general circulation model (GCM; past/future climate) climatologies. The ratio of reanalysis to past or future climate depends on a user-selected global temperature departure value, ∆T. For the LGM, ∆T = -6 °C; for modern climate, ∆T = 0 °C; for 2100, ∆T = +4 °C. And so on.
This version of the ECM is adapted from Birkel, 2010 (PhD. Dissertation).
How do I use the ECM?
View different ECM solutions by selecting a Region, Parameter, or Month from the pulldown menus, and by hovering the cursor over ∆T values. Use the Playback controls to animate the images. The arrow and stop buttons control frame advance; the "-" and "+" buttons control frame speed. If images do not update on playback, then you likely have a low bandwidth network connection. In that case, click "-" to slow down the playback, which will give images more time to load.
What else should I know?
Model Input Climatologies
- NASA MERRA (Reanalysis - modern)
- 1979-2000 average. Datafile source.
- NCAR CCSM4 (GCM - past/future)
- LGM (simulation b40.lgm21ka.1deg.002; 50-year average). Datafile source. See also Brady et al., 2013.
- 2100 (simulation RCP8.5 [high industrial emissions]; 2095-2100 average of 6 ensemble members). Datafile source.
Model Grid Resolution and Temperature Downscaling
The ECM produces solutions on 2.5, 5, or 10 arcminute grids (i.e., ~ 4, 8, or 16 km) depending on the selected domain and size of the output image. Neither the reanalysis (~60 km) or GCM (~120 km) input grids provide this high resolution, therefore a statistical downscale method is employed linking temperature to topographic elevation:
- ETOPO1 (1 arcminute) global topography is obtained and regridded to 2.5, 5, and 10 arcminute resolution files.
- LGM - modern and 2100 - modern difference grids are calculated at the original coarse GCM resolution, and then resized using bilinear interpolation to overlay the ETOPO1 files.
- NOTE: In this step, "modern" refers to the modern control run of the GCM. Differencing is done within the GCM for internal consistency.
- The coarse resolution reanalysis input grids are resized using bilinear interpolation to overlay the ETOPO1 files. Monthly temperature fields are downscaled to the resolution of regridded ETOPO1 topography using a 5 °C/km atmospheric lapse rate.
- The GCM-modern difference grids calculated in Step 2 are then applied to the resized reanalysis grids produced in Step 3.
Reanalysis-GCM Blending
In the real world, Earth's mean climate state evolves in non-linear dynamical response (e.g., see detailed description of dynamical systems) to radiation imbalances arising from a variety of driving stresses. Natural driving stresses include changes in solar output, orbital geometry, and volcanic activity; human driving stresses include changes in land use (alteration of surface albedo), and emissions of greenhouse gases and ozone-depleting compounds. The ECM does NOT compute the dynamical response of the climate system. Instead, the ECM uses output fields from reanalysis and general circulation models that have already done the heavy numerical computing, and blends between selected climate end members.
MERRA (reanalysis) defines the modern climate end member; CCSM4 (GCM) defines the LGM and 2100 climate end members. Adjusting the global ∆T in ECM changes the blending ratio in equivalent steps between MERRA and CCSM4. For example, ∆T = 0 °C affords a blending 100% MERRA and 0% CCSM4, whereas ∆T = +2 °C affords a blending 50% MERRA and 50% CCSM4 2100; ∆T = +4 °C affords 100% CCSM4 2100. Likewise for cooling simulations from ∆T -0.5 to -6 °C, where the latter is 100% CCSM4 LGM.
Biome Rubric
A biome is a region defined by characteristic assemblages of plants and animals. There are several existing biome models published in the scientific literature (e.g., Nemani and Running, 1996; Kaplan et al., 2003), each with different solution algorithms and input datasets. The ECM uses a hybrid biome classification scheme derived from multiple sources, and from examination of satellite land surface imagery in Google Earth in relation to reanalysis climate fields. The ECM method first establishes four primary biome types based on total annual precipitation (humid, intermediate, semi-arid, desert), and then proceeds to further classify based on temperature of the warmest and coldest months. A snow/ice mass balance calculation is also made at each gridcell in order to determine areas of perennial snow or glacier ice. The ECM biome classification scheme follows, with temperature values in °C and precipitation in meters:
| Primary Biomes | ||
| Humid | = | P_annual >= 1.5 |
| Intermediate | = | P_annual >= 0.62 AND P_annual < 1.5 |
| Semi-arid | = | P_annual >= 0.25 AND P_annual < 0.62 |
| Desert | = | P_annual >= 0 AND P_annual < 0.25 |
Humid Biomes | ||
| Tropical Rainforest | = | T_summer >= 21 |
| Temperate Rainforest | = | T_summer >= 10 AND T_summer < 21 |
Intermediate Biomes | ||
| Broadleaf Forest | = | T_summer >= 23 |
| Mixed Boreal-Broadleaf Forest | = | T_summer >= 18 AND T_summer < 23 |
| Boreal Forest | = | T_summer >= 11 AND T_summer < 18 |
| Moist Tundra or Alpine | = | T_summer >= 0 AND T_summer < 11 |
Semi-arid Biomes | ||
| Tall Grass Prairie | = | T_summer > 18.5 AND P_annual > 0.45 |
| Short Grass Prairie | = | T_summer > 18.5 AND P_annual > 0.35 AND P_annual < 0.45 |
| Steppe | = | T_summer > 18.5 AND P_annual < 0.35 |
| Cool Steppe | = | T_summer > 12 AND T_summer < 18.5 AND T_annual > 1 |
| Boreal Forest | = | T_summer > 13 AND T_summer < 18.5 AND T_annual < 1 |
| Dry Tundra or Alpine | = | T_summer < 13 |
| Forest-Tundra Transition | = | T_summer >= 10 AND T_summer < 13 AND T_annual < -1 |
Desert Biomes | ||
| Polar Desert | = | T_annual <= 0 |
| Low Latitude Desert | = | T_annual > 0 |
Sea Level
From LGM (∆T = -6 °C) to modern climate (∆T = -0 °C), the sea-level datum (SL) increase exponentially from -120 m to 0 m following the equation ∆SL = (dT^2)/(-36)*120. This simple, non-physical relationship is used to represent vast changes sea level linked to the size of ice sheets over North America, Europe, and Asia. For future climate scenarios, we define ∆SL as +0.25 m (∆T = +1 °C), +0.5 m (∆T = +2 °C), +1.5 m (∆T = +3 °C), and +1.75 (∆T = +4 °C) in accord with assumed ice loss from the Greenland Ice Sheet. End member ∆SL values (i.e., for LGM and 2100 CE) are commonly reported in scientific literature. Bear in mind that the ECM does NOT calculate ice volume or sea level in a physical way.
Interpreting ECM Solutions
The ECM is a model, and, as with any model, its output is limited by input datasets and assumptions that underlie the model. The input datasets in this case are derived from reanalysis and GCMs. They too are limited, but shown in peer review literature to provide reasonable solutions for the state of Earth's climate at different time intervals. With this in mind, ECM users are encouraged to evaluate the potential biome solutions. Does the predicted distribution of biomes under modern climate conditions make sense? Are there deserts where there should be deserts? Does tundra map where there should be tundra? Both success and failure of a model can yield useful information about how a complex natural system operates in reality.
This material is based upon work supported by the National Science Foundation under Grant Number 1107421. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.
