Laser Trackers (LTs) are useful metrology tools for quickly acquiring accurate 3D position measurements of a target over long ranges (e.g., up to 100 m). These tools often employ an Absolute Distance Meter (ADM) which is known to have a temporal measurement shift arising from the internal thermal effect. The measured radial distance of a regular target can be compensated by subtracting from it the ADM shift. However, for many LTs currently in service the ADM shift cannot be acquired when an LT is in Operation. In this work, an estimate of the ADM shift is adopted for radial distance compensation when an LT is in Operation. A Switched Estimation (SE) algorithm is proposed to evaluate the validity of the estimate. The SE algorithm leverages the estimation uncertainty given by a Modified Kalman Filter (MKF) and confines the estimation uncertainty within a pair of predefined boundaries. When the estimation uncertainty reaches the upper bound, a regular Kalman Filter (KF) is adopted to calibrate the ADM shift estimate, during which the LT is directed to measure the ADM shift. Once the uncertainty reaches a lower bound, the LT is redirected to be back in Operation. Obviously high estimation accuracy, consequently high compensation accuracy, conflicts with high operation productivity. A tradeoff can be made by selecting an appropriate pair of boundaries. A numerical method is given for the selection of proper boundaries by creating a selection map. Experimental results show that with the proposed SE algorithm, the maximum variation of the radial distance measurement is reduced by more than 70% while the operation productivity is kept at 93.02%.