…i’ve only done urban installations, but recently i’ve been giving a lot of thought to the localised effect of high-albedo reflective roofs (and other materials) and transpirative tree canopies (and other vegetation) being replaced by low-albedo solar photovoltaic arrays: it measurably increases heat load on the local environment, reduces radiative cooling at night, and drives up overall cooling demand, which presents a deep rabbit-hole of cost-benefit analyses in the tradeoff between reduced grid use during the day versus increased grid use at night, the potential net reduction in carbon emissions therefrom, the added carbon emissions from manufacturing and maintaining solar photovoltaic infrastructure, and the loss of ecosystem carbon capture and biodiversity services from decreased solar exposure and increased heat island effect…

…it’s a poorly-understood subject of ongoing academic study in both urban and natural environments, with the largest arrays i’ve read subjected to that sort of rigorous analysis measuring on the order of 1/300,000 the scale of this proposed project…still, apples-to-apples, that larger study was performed in desert scrubland and measured about 4°C increased local temperatures, which is significant but not a good proxy for the weather effects one would see generated by a 750 square-mile convection cell over truly barren desert…

…back to your original question, most solar photovoltaic panels loose somewhere on the order of ¼ to ½ percent efficiency per °C incease in panel temperature, but like most things in the real world it’s actually much more complicated than a simple multiplier…the short version is that investors wouldn’t be building desert arrays if they didn’t present a short-term economic gain, and they certainly do provide plenty of power despite the increased heat, but the long-term environmental impact of radically altering surface albedo at such a large scale isn’t well-understood relative to the implied let alone actual changes in carbon-intensive energy generation…