Release notes (2020-08-20)
Version 2.1 provides a fix for present-day current velocity data and some minor oddities occurring in certain future projections (see technical note for details). We encourage all users to use this updated version.
Present climatic conditions: - Surface: Currents velocity - Benthic: Currents velocity
Future climatic conditions [Representative Concentration Pathways, 2050 and 2100:
- Surface: All layers
- Benthic: All layers
Note that, in order to keep versioning of layers consistent, we provide a full set of version 2.1 layers, but only the layers listed above differ from version 2.0. We keep version 2.0 available to the community to facilitate reproducibility of research.
1. Layers for present conditions
For present-day conditions, version 2.1 corrects the allocation of variability found in the temporal data used to produce the predictor layers of 'currents velocity'.
The previous version 2.0 determined the velocity of ocean currents using Pythagoras' theorem over the northward and eastward components of ocean velocity fields. In this process, the predictors 'minimum', 'maximum', 'long-term minimum' and 'long-term maximum' of both ocean components were stacked and paired to determine the corresponding layers of 'currents velocity', by considering their actual values. These are found positive to define northward and eastward current velocities and negative to define southward and westward velocities. The use of this information to determine 'currents velocity' should use the absolute values of both components and neglect the relative direction of currents. For instance, the previous layer defining the minimum northward component might have included strong southward currents, when values were negative and extreme. The new layers of 'currents velocity' in version 2.1 were calculated using the absolute values of both components in Pythagoras' theorem. This way, the predictors 'minimum', 'maximum', 'long-term minimum' and 'long-term maximum' reflect the actual intensity of currents, regardless of their relative direction.
2. Layers for future conditions [Representative Concentration Pathways]
For future conditions under the different RCP scenarios, version 2.1 corrects the allocation of variability found in the temporal data used to produce all predictor layers.
Figure 1. Change-factor approach using the projected magnitude of climate change inferred from Ocean General Circulation Models (AOGCM). Example of change estimated for the decade 2090-2100. Bold depicts the final set of layers available in Bio-ORACLE.
Future conditions in Bio-ORACLE are determined using the state-of-the-art change-factor approach (see methods sections; Assis et al., 2017). This is based on applying the projected magnitude of climate change, as inferred from different Atmospheric Ocean General Circulation Models (AOGCM), to the layers available for present-day conditions. Data for a baseline period (e.g., 2000–2014) is obtained from the AOGCMs to determine the difference (change-factor) between the present conditions and the future scenarios of change. Later, the change-factor is applied to the layers with present-day conditions to obtain the final RCP climate estimates (Fig. 1).
In this process, it is possible to obtain predictor cells where minimum values exceed maximum values when the following conditions are met:
a. The difference in between 'maximum' and 'minimum' for present-day conditions is very small (e.g., temperature or salinity at the offshore bottom of the ocean, be low 1000m depth).
b. The change-factor projects bigger changes for 'minimum' than 'maximum' predictors. This is possible when addressing the magnitude of climate change. For instance, temperature might have higher projected changes in colder months than in warmer months.
By applying a higher change-factor to 'minimum' predictor cells with residual variance in present-day conditions, it is possible to end up with 'minimum' > 'maximum' cells in the final RCP layers. The new layers in version 2.1 for future conditions re-allocate minimum and maximum values to the corresponding layers representing 'minimum' and 'maximum' predictors, resulting in an accurate representation of the projected climate variability while retaining the logical order of minimum and maximum.