You're arguing in good faith. Which, good, don't let them take that earnest attitude. But she is not.

She is getting paid to spread disinformation. Whatever actual misconceptions were there to begin, are probably now tied with deception and inseparable. Getting paid to lie makes it really easy to lie to yourself too.

Relevant open access research on this

https://www.sciencedirect.com/science/article/pii/S0306261926006604

Most studies, however, tried to capture direct impacts of solar PV sites on biodiversity through field measurements, but even within these, methodological approaches varied considerably. For example, while many studies used the surrounding landscape as the control site [e.g. [56], [59], [65], [75]], others assessed biodiversity only within solar PV sites, using areas between panels or next to panels as control [e.g. [77], [91], [92]. Although such within-site comparison can provide useful insights into the effects of shading or altered water patterns on soil organisms or plants, from a life cycle perspective it is desirable to take the nearby (semi-)natural environment as a control environment to be able to make conclusions on the potential loss of biodiversity. An alternative approach would be to measure biodiversity levels before installation and again some years after [e.g. 104].

Additionally, while it is important to understand the benefits solar PV sites can provide compared to, for example, intensely managed agricultural sites [e.g. [75], there is still the need for a standardised control condition across biodiversity assessments. For this reason, LCIA methods typically use a (semi-)natural reference landscape as the control condition, ensuring that the impacts of any land-use type can be assessed against a common baseline.

https://www.sciencedirect.com/science/article/pii/S2666683926000787

also relevant https://www.pv-magazine.com/2026/05/18/spain-france-set-benchmark-in-renewables-expansion-without-biodiversity-loss/