if the transceivers are positioned statically (not moving), then choose a directional antenna and quite small output powers will provide you a stable connection. But be notified that the 2.4 Ghz is severed by weather quite a lot.
If not and the transceivers are moving, let's make some path loss calculations:
transmit power in dBm - path loss in dB > receiver sensitivity in dBm
According to the datasheet, the RX sensivity at BER=0.1% is -80dBm.
The path loss is the actual path loss + transmitter antenna gain + reciver antenna gain. For this simple equation, we assume that both antennas are 0 dBm.
The path loss at 2450 Mhz for 1 km in air is 100dB (this is quite nice calculator for it: www.qsl.net/.../jsffield.htm )
transmit power - 100 > -80
transmit power > 20dBm
In real situation, please notify that this is only a theroretical value. Terrain and bad weather will severe the communication. For proper data transmission, I think you need about 30 dBm output power. For this, use a RF power amplifier.
From my experience, putting a LNA to the receiver won't make the range better a lot. You will amplify the signal let's say about 20dB, but the noise as well (and the LNA will add another several dB of noise to the output). I have tested this at 868 Mhz and I expanded the range about 10%. To be able to reconstruct the signal, the S/N ratio must be over 1. If the signal goes below the noise floor, it is generally lost (not really, but here I am not talking about lock-in amplifiers and other complex and expensive stuff).
However, running an RF device at so high output powers probably will be a violation of FCC and/or ETSI rules for unlicensed operation. Take a look at the rules. If your device will violate the regulations, use a solution with lower carrier frequency (path loss is smaller at lower frequencies, at 868 Mhz @ 1km is about 90dB and at 433 Mhz @ 1km about 85 dB). The power limits at 868 Mhz and 433 Mhz are also slightly more relaxed I think.
