Review
This manuscript addresses a key concern of our time: managing landscapes that should best fulfill ecological (explicitly wildlife needs) and economic goals (farmers' needs). Future sustainable land use will critically depend on innovative tools and approaches to analyze and understand the resulting trade-offs and how to mitigate them. I furthermore highly value the open-source code approach used here. I therefore believe that this is an important contribution. The manuscript itself should, however, be carefully revised in order to clarify some crucial aspects outlined below.

Major points:
1. The section on existing model approaches (lines 55-62) and, thus, the framework for the innovation of the model used here is, in my opinion, not complete enough. There are various approaches in Multiobjective Optimization that combine ecological and economical goals where the objective is not the maximization of all goals simultaneously (e.g., threshold or reference point approaches). Showing further models can help to work out the innovation of the manuscript more clearly. The unique part is the combination of GA and a process-based model, exemplified with wild bees. I think streamlining the introduction more by clarifying which models exist and how the here used approach is innovative.

1. In combination with 1. the structure of the introduction is not entirely straightforward. After the very first sentence, it goes straight to methodological aspects, after which it first deals with wild bees and their relevance to the major problem of global biodiversity loss, then comes back to the fact that the combination of multi-objective optimization and process-based ecological models is the innovation of the manuscript. I would suggest first outlining the problem in more detail, followed by an extended section of available models. This would allow to clearly state a research gap. Connected to this, I would also recommend to formulate clear research questions. The objectives of the paper are clearly stated but deriving clear research questions that could be (exemplarily) solved with the suggested method would improve readability and help to outline the novelty and relevance of the approach

2. The optimization goals are not explicitly mentioned in the introduction. Later, it becomes clear that the objective function is to increase floral visitation rates/abundance and income (without risk and costs). The optimization goals could be directly mentioned in the introduction (maybe in lines 108-115). I personally would appreciate a mathematical description of the objective function for reproducibility (maybe in a reduced form in the appendix). Going through the entire code is maybe too demanding to understand the methodological novelty. Also, the aspect of how the approaches deals with uncertainty could be described in more detail.

3. Definitions of spatial explicitness: Since the distribution of land use in space is of great relevance, it is important to define the characteristics of the spatial explicitness used. If I understand it correctly, spatial explicitness consists of mean patch area and patch location. So, there is, for example, no explicit inclusion of landscape heterogeneities (e.g. site differences etc.) and connectivity? Some clarification in the introduction concerning the conceptual understanding and additional information in the method section would be helpful. Also, the spatial scale of the original dataset is not fully clear.

Minor points:

1. Land cover/land use and landscape are occasionally used interchangeably (e.g., lines 171-175). The author should explain whether the terms are the same or how they are defined.

2. In line 179, what is the defined number, and how is it defined?

3. In line 189, what is the field boundary template?

4. In Table 1, what does the abbreviation AUK mean? Agriculture UK? It should be clarified in the caption.

5. The equal weighting of all objectives (line 276) means that the overall needs of the bees are weighted more heavily than the needs of the farmer, as the optimization includes three bee objectives and one farmer objective. Even though there are three different bee guilds, it might be worth acknowledging in the discussion.

6. Specifying p-values (line 295) is increasingly discussed. Specifying the effect size with values, e.g.., which percentage difference between the land use areas, is more important.

7. Providing the average percentage of land cover types for the "real landscape" would be a helpful addition in Figure 3. For example, one could see whether the solution landscape for farmers is quite close to the "real landscape".

8. Related to 12., I find it quite counterintuitive that the solution landscape for the farmer was lower than in the real landscape (lines 342-344). Since the manuscript states that other approaches may not reflect a true potential optimum, the author should directly address why this is not the case for this farmer result. Section 4.2.3 explains this in more detail, but this is quite late. Showing the land cover percentages in Figure 3 could help to address this.

9. Can the authors' method be applied to other combinations of GA and process-based models? Are there examples of other models that could follow the same approach?

10. Please check the capitalization of article titles and journal names. Currently, the use of upper and lower case is not uniform (e.g., lines 636-638).