[Bug] timestamp reading from paimon is different from hive after hive-cloning

### Search before asking

- [x] I searched in the [issues](https://github.com/apache/paimon/issues) and found nothing similar.

### Paimon version

Paimon's version: 1.2-snapshot(2025/06/09)

### Compute Engine

Hive's version: hive 2.3.10

Flink's version: 1.20.1

### Minimal reproduce step

When selecting the paimon table which was cloned from hive through hive-clone action, the selected result for timestamp is different between paimon and hive.

**Note: file format is parquet.**

Minimal reproduce step:
1. create a table in hive and insert some data:
```sql
-- create a table
CREATE TABLE test.hive_test_table(
`a` int COMMENT 'The a field',
`ts` timestamp COMMENT 'The ts field'
) STORED AS PARQUET;

-- insert a value
INSERT INTO test.hive_test_table VALUES (1, '2025-06-03 16:00:00');

-- select
SELECT * FROM test.hive_test_table;
-- result: 1, 2025-06-03T16:00
```

2. run hive-clone action

3. select the cloned paimon table in flink
```sql
SELECT * FROM paimon_hive_test_table;
-- result: 1, 2025-06-03T08:00
```

The reaseon may be:
Hive will transform the timestamp to UTC for storage, and transform the stored timestamp to `writer.time.zone` or system default time zone when reading. While paimon will directly read the original timestamp stored in parquet file without any transforms, this leads to difference.

### What doesn't meet your expectations?

The selected result for timestamp has changed after cloning to paimon.

### Anything else?

_No response_

### Are you willing to submit a PR?

- [ ] I'm willing to submit a PR!

Run the listed Hive SQL setup, hive-clone action, and Flink query using the stated versions and Parquet format. Compare the timestamp before and after cloning, then trace the clone and timestamp-reading paths to determine where timezone handling diverges. Done means the cloned Paimon query returns the same timestamp as Hive.

Written by the indexing model from the issue text.