Wow...That's a problem! I would try the following method. First get a REAL GOOD topographic map of your exact location. Note the latitude of your 'observatory' to arc minute accuracy. Then, with a compass you will have to identify a landmark on your horizon that is at magnetic due north. Then, from the map, correct this by the local magnetic deviation to find true north. The landmark on your local horizon in this direction should then be true north. Then, from your latitude, subtract it from 90 degrees. Using this angle, this gives the number of degrees above the horizon above your true north landmark that the North Celestial Pole is located. With a theodolite and a good compass, you could get your polar axis boresighted to within a few arc minutes this way. The HARDEST STEP is measuring your magnetic north to less than 1 degree accuracy, and this will limit just how good your initial guess for the true celestial pole will be.
Next, you will have to select several stars at high declination. Center the star in your eyepiece by rotating the polar and declination axis, without moving the mounting. Now, you have to shift the telescope pointing by one eyepiece field in Right Ascension and see if the star drifts back to its center position. The amount of deviation tells you how far off from being On The Pole, your polar axis is. Now, the elevation angle of your mounting should be pretty accurately at the right angle if you were careful, but the pointing to the north pole using the magnetic compass is probably the least accurate so I would start by slightly shifting the mounting rather than by adjusting the polar axis tilt angle. By trial and error, you may get the alignment you need for most, short term, observing sessions. It is a difficult process, even under the best of circumstances!