No special difficulty was encountered during the fourth and last year of the ALATNET network. The completion of the work programme, the quasi achievement of the manpower target (1047.5 against the initially foreseen 1055 person × months) and the meaningful spending of the remaining budget went on without any hiatus. Once the agreement of the TMR programme was obtained for the slight shift (at constant envelope) of budgetary amounts necessary to organise it, the newly-added Kiralyret seminar for the young researchers was prepared and took place also without any problem, thanks to the organisation efforts of the Hungarian Partner.

On the financial side, the foreseen budget of 822000 euros has been spent at a relatively controlled pace over the four years of the project (10.5% / 33.0% / 33.7% / 22.8%, the contract foreseeing 21.9% / 33.3% / 27.0% / 17.8%) while keeping the total overheads below their 13% target (by about 2000 euros). The detailed spending went roughly according to proposals (with an average 3.3% relative error on final individual lines of expenses when compared to the total sum). The part of the budget spent in support to young researchers activities amounts to 61.1%, above the 60% deliverable requirement.

This question is partly irrelevant since Numerical Weather Prediction is a kind of "auto-industry" at the heart of the very complex system of data collection, treatment and redistribution for weather-linked routine and/or extreme events (see Part B.5.7).

Nevertheless, the above-mentioned complexity of the system in which ALATNET actions were embedded offers an interesting way to look at the Network achievements : did it influence the evolution of that specific super-structure or did it remain an independent research branch of it ? Seen from this angle the response is certainly that of the first alternative. ALATNET has marked the evolution of NWP at the level of links between the various European groups specialised on the same topics (and hence more generally the meteorological collaboration in Europe). This is well synthesised A) by the decision to take the ALADIN NH-dynamics and 3d-var "bricks" to build the new generation NWP system of Météo-France and of the ALADIN Partners and B) by the now confirmed choice of the HIRLAM group to join this already ongoing new development around tools developed thanks to ALATNET.

Since the ALATNET team is very pleased about the way things went on and about the communications with the TMR responsible people in DG-XII, we do not have too many suggestions for improvement.

We would have preferred that the reporting procedure be fixed for the whole length of the contract.

We would have liked to have an official formula allowing to foresee the intermediate payments from the rate of consumption of the ALATNET money. For Services without too much cash flow, like Partners 3, 4 and 5, it would have been helpful to know already in April roughly which sum they would get reimbursed in the Fall.

One other problem is the non-uniform (not very Euro-like thus) treatment of the Young Researchers' remuneration : the taxation in Belgium introduced a distortion (already known at the time of negotiating the contract) in the split between part A, B and C financial resources and, more annoyingly, created a potential dependency of the spending on changes of social legislation in the country during the length of the contract.

B6-8-2