The time factor and how it plays out
I think @Nadia and @Hazem are very right to highlight the time factor.
The way I see it: for the WB project, we need some quick, easy win to show. Remember the unMonastery crowd's reluctance to commit to results within the duration of the prototype, insisting the time frame was unrealistic. This rules agriculture right out: by definition, it follows the rythm of the seasons. Most crops need a minimum of one year. If you are looking at olive orchards, you are going to need twenty.
On the other hand, Tarfaya or some other out-of-the-way place has advantages for the WB-funded prototype, some of which others have already pointed out. If the unMonastery experience is anything to go by, a Reef in Tarfaya will be highly conspicuous – maybe even too conspiscuous, but that's a good problem to have. However, out-of-the-way locations are hard to reach for people in the region, and sometimes even in the country, as Hazem points out.
BUT if we are looking at an ER long-term investment, the case for cities (real cities, @matthias , not 50K inhabitants towns) becomes much stronger. If we are to invest, we need sustainability, and we need it right away, because we do not want to go into debt. For now, the only proven source of revenue in the region is consulting on development issues. I am persuaded that having a presence in the region might be an advantage, allowing us to get more contracts than we do now. Also, it could become a template for future expansion: we could spawn a Reef hosting Edgeryders West Africa, Edgeryders SouthEast Asia or whatever. But in this case, we would need to climb the learning curve of getting these contracts, and I am afraid that calls for being close to clients and donors. Also, we need a place where we can ourselves go quite quickly and cheaply as needed. In Morocco, that means to a first approximation Rabat (but Casablanca is close enough that you can get away with it). As you get established, you can probably get away with relocation, but in the early days the economics of a space far from sources of money are the economics of fixed costs and burnthrough rates. That's quite risky.
So, here's a suggestion: maybe decide if we are prioritising doing a great prototype now, and then maybe we'll make a permanent investment. In this case we are more free to pick locations, because the prototype is already funded. On the other hand, if we insist that the location of the prototype be ready to stabilise, then we need to be conservative and prioritise the ease of making sales – which is the hardest bit of the ER experience.
In all cases, the idea of crowdsourcing locations is great.