[Data Issue]: Possible double counting of storage related emissions in consumption view

[leon-hard]

When did this happen?

Ongoing

What zones are affected?

all

What is the problem?

I have a problem to understand how double counting for storage related emissions is avoided at EM. I don't want to discuss here if it is reasonable to take an average value for the emissions related to the grid of the respective country of last year, whereas we know that the business model of these pumped hydro plants is to store variable production, instead to understand how the load for each country is calculated.

I was always assuming that EM is using the load as defined by ENTSO-E without pumped storage consumption for flow tracing and emission calculation, which would therefore avoid double counting. I defended EM already various time if people raised the question on double counting. However, if I have a close look at the data available at the interface, I have more the impression that the load is calculated as production + import saldo. But this would lead inherently to double counting. I could not find any clear defintion how you bring production and consumption in balance for your flow-tracing implementation, which is crucial to understand storage related emission counting.

Find attached the graph that somehow let's me assume that the load is calculated as production + import saldo.

I would be happy to get more information on the calculation of the load, maybe it is also just already documented somewhere and I was to stupid to find it.

Here a simple rational of my thinking:

1. Plant A produces 2 kWh electricity with emissions of 35 gCO2eq/kWh
2. 50% are stored (1 kWh), 50% are consumed (1 kWh) -> these two 2 kWh production are then accounted to load (defined as total production + import saldo for each country, therefore included pumped hydro consumption)
3. In a later timestep, the pumped hydro is producting electricity with 35 gCO2eq/kWh. The load by then will again account for this production related emissions.

But like mentioned above, this all boils down to the question how the load is calculated.