There is always ice.
Secondly, where are you located?
In India, they use a higher temperature (27degC) for the tank. I will attach a relevant part of Indian Standards - although 1959 edition in a "reply" to this reply as the site doesn't permit two attachments in one reply, I remember that this was about right when I was working there in 2001 to 2005.
Do you have a nearby river source or water source? if so, what is its temperature. We find, even in the hot climate of Malaysia that the river temperature is fairly constant at about 25degC even in the summer - if this is so, you can maintain your temperature by continual circulation. Or, as it most usually the case, do nothing. I am presuming that you are describing a site lab for a construction project. I know a very respected concrete expert who has been trying to convince his peers that for site labs - live with the ambient temperatures (within reason). I have attached a figure from Neville's book Properties of Concrete which shows the differences in % of 23degC curing temperature and other curing temperatures at various dates. It seems that at 90 deg F one is about 95% of the "control" strength at 90 days; at 45 degF about 110%. What you might wish to do, too, is to put in a curing tank of some volume and see what the actual temperatures can be maintained at - the 100degF must be a daytime temperature but you will have a lower, perhaps 75 degF temperature at night so the actual ambient might be 85degF (27 degC - similar to what is used in India).
You could always do some cubes (or cylinders) and cure in ambient at your lab and properly in the compliant lab and see what difference actually happens. Also, the effects of the age of test would be likely more relevant (Ron correct me if I am wrong) - some jurisdictions I have dealt with over the last 8 years specify a 90 day test for compliance given that fly ash is used in the mix.
