Paper: arXiv 2310.10518
Authors: Nandi Moksnes, William Usher
Abstract
An increasing number of studies using energy system optimisation models are conducted with higher spatial and temporal resolution. This comes with a computational cost which places a limit on the size, complexity, and detail of the model. In this paper, we explore the relative importance of structural aspects of energy system models, spatial and temporal resolution, compared to uncertainties in input parameters such as final energy demand, discount rate and capital costs. We use global sensitivity analysis to uncover these interactions for two developing countries, Kenya, and Benin, which still lack universal access to electricity. We find that temporal resolution has a high influence on all assessed results parameters, and spatial resolution has a significant influence on the expansion of distribution lines to the unelectrified population. The larger overall influence of temporal resolution indicates that this should be prioritised compared to spatial resolution.
Complexity vs Empirical Score
- Math Complexity: 6.0/10
- Empirical Rigor: 7.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced global sensitivity analysis (e.g., Sobol indices) with non-linear model interactions, and demonstrates empirical rigor through case studies in Kenya and Benin with detailed parameter exploration and error bars.
Research Flowchart
flowchart TD
A["Research Goal<br>Determine relative importance of<br>spatial vs. temporal resolution<br>in energy system optimization models"] --> B
subgraph B ["Methodology & Data"]
direction TB
B1["Global Sensitivity Analysis GSA"]
B2["Input Data<br>Energy demand, Discount rate,<br>Capital costs, Grid constraints"]
B1 & B2 --> B3["Model Application<br>Kenya & Benin case studies"]
end
B3 --> C["Computational Process<br>Iterative simulation varying<br>spatial & temporal resolution<br>with parameter uncertainties"]
C --> D["Key Findings Outcomes"]
subgraph D
direction LR
D1["Temporal Resolution<br>High influence on all results"]
D2["Spatial Resolution<br>Significant influence on<br>distribution line expansion"]
end
D --> E["Conclusion<br>Prioritize temporal resolution<br>over spatial resolution"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes that temporal and spatial resolution should be customizable and possibly automated when modelling network expansion, depending on the research question.
Source sentence
The paper recommends that great care is needed when formulating the temporal resolution as this affects the overall insights of the modelling.
Assumes 6
- A uniform discount rate is used for all technologies. model
Source sentence
In this study we use a uniform discount rate, however, a technology-specific discount rate could alter the results, and thus would also be a suitable future parameter to include.
- The demand profile for the electrified population is assumed identical for Kenya and Benin and constant over time. data
Source sentence
The demand profile for the electrified demand was assumed to have the same profile for both countries, and it remained the same throughout the modelling.
- The model assumes linearity in the representation of transmission and distribution costs. model
Source sentence
Furthermore, due to the linearity assumption, very small amounts of transmission lines can be installed at quite a low cost.
- Spatial resolution is varied programmatically but the study area is divided as evenly as possible. execution
Source sentence
Furthermore, the spatial resolution in this study was based on making the area as even as possible.
Show 2 more
- The temporal resolution is represented by discrete time slices ranging from 1 to 24 per day. model
Source sentence
The temporal resolution varies from 1 to 24 timesteps per day (i.e., hourly at the highest granularity), and as illustrated in Figure 4, one can see that the average value of one time slice is flat throughout the day while the 24 time slices follow the original data set most closely. Figure 4.
- No seasonal representation is included in the GSA runs. model
Source sentence
In this GSA, there was no seasonal representation which can influence the results where there are seasonal changes e.g., hydrology, wind power, demand.
Data
- GEOSeMOSYS synthetic public
Source sentence
For this paper, we develop the open-source code of GEOSeMOSYS further to also include varying temporal resolution, demand, discount rate, fuel, capacity factor error, and capital cost to run a GSA (see Table 1).
- OSeMOSYS synthetic public
Source sentence
Country model data To address the paper’s research question we use the open-source ESOM OSeMOSYS 36.
- Source code repository GSA_Spatial_temporal other public
Source sentence
Data and Code Availability The source code used in this study is available at GSA_Spatial_temporal: https://github.com/KTH-dESA/GSA_Spatial_temporal.
- TEMBA data for Benin electricity sector fundamentals public
Source sentence
The general powerplant structure and future potentials, which is a pre-requisite to the GEOSeMOSYS model generator, are extracted from TEMBA39 for the Benin electricity sector (excluding trade with neighbouring countries) and Kenya the model from the paper29.
Costs
Not applicable
Finds 5
- Temporal resolution has a high influence on all assessed result parameters and is more influential than spatial resolution. theoretical
Source sentence
We find that temporal resolution has a high influence on all assessed results parameters, and spatial resolution has a significant influence on the expansion of distribution lines to the unelectrified population. The larger overall influence of temporal resolution indicates that this should be prioritised compared to spatial resolution.
- Spatial resolution significantly influences the expansion of distribution lines to the unelectrified population. theoretical
Source sentence
We find that temporal resolution has a high influence on all assessed results parameters, and spatial resolution has a significant influence on the expansion of distribution lines to the unelectrified population. Spatial resolution has a high influence on the distribution line's expansion, but a more moderate impact on the total cost and renewable electricity production.
- Demand and discount rates are highly influential factors in the model outcomes. theoretical
Source sentence
The demand is highly influential, and the levels of Tier 1 to Tier 4 for the unelectrified population in the base year is a large range. Apart from the structural parameters, we can also see that demand and discount rates are highly influential, and a more detailed representation of these can better help us understand the evolution of the system.
- Proper representation of the demand peak is important; omitting the peak reduces total system cost. theoretical
Source sentence
Second, the temporal resolution of one time slice omits the peak, leading the demand to be the same as the flat value (as illustrated in Figure 3A with the brown dashed line (average PV capacity factor) and black dashed line (demand profile)). When the peak is omitted, the total cost of the system is reduced as in OSeMOSYS the controlling factor for cost of capital is the modelled peak demand.
- The method of Morris provides a computationally efficient factor prioritisation but has limitations for non‑influential parameters. theoretical
Source sentence
The method of Morris is a computationally efficient method to understand sensitivities in a model. This highlights the limitation of the method of Morris, which can only give indications of influences, especially to non-influential parameters.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 21 related papers). Coverage is partial, so this is neither replication nor refutation.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.