Transshipment nodes are the least substitutable components of a multimodal freight network: rail and barge cargo cannot detour around a closed terminal the way a truck detours around a closed road, so how a model represents node failure determines whether it captures the network's real point of weakness. Existing macroscopic models nevertheless treat a disrupted node as either fully open or fully closed, which misses the majority of real events, since more than half of observed port disruptions are partial. This paper presents a day-to-day simulation framework for a transshipment node. It applies continuous, mode-specific capacity loss, lets unserved demand accumulate as a physical queue that rolls over between days, and releases only a limited share of that queue to alternative routes each day, since real shippers cannot re-plan overnight. The framework is applied to Duisburg, Europe's largest trimodal inland port, using the European TEN-T network and empirical European freight demand data; two further nodes of different scale and modal composition confirm the pattern generalises. Partial degradation is not simply a smaller version of complete failure. A 50% capacity loss produces roughly a quarter of the full-closure cost. A binary model treating a half-degraded node as closed therefore overstates its cost by a factor of about four, while one treating it as operational misses that cost entirely. More importantly, what actually limits recovery is not a shortage of alternative routes but how quickly freight is contractually free to use them. Alternative nodes absorb diverted cargo from day one, yet removing the re-planning constraint entirely cuts the cost of a 51-day closure of Duisburg from €2,887.6M to €1,095.5M. Re-planning inertia alone therefore multiplies the damage by about 2.6. Contractual flexibility, not only physical capacity, belongs in resilience planning. ... Read more