What you have here is an example of the concept of memory-mapped peripherals. Basically, the MCU hardware has special locations in the SRAM address space of the MCU assigned to it. If you write to these addresses, the bits of the byte written to address n control the behaviour of peripheral m.
Basically, certain banks of memory literally have little wires running from the SRAM cell to the hardware. If you write a "1" to this bit in that byte, it sets this SRAM cell to a logical high, which then turns on some portion of the hardware.
If you look into the headers for the MCU, there are great big tables of keyword<->address mappings. This is how things like TCCR1B etc... are resolved at compile time.
This memory-mapping mechanism is extremely broadly used in MCUs. The ATmega MCU in the arduino use it, as do PIC, ARM, MSP430, STM32 and STM8 MCU series, as well as lots of MCUs I'm not immediately familiar with.
Arduino code is the weird stuff, with functions that access the MCU control registers indirectly. While this is somewhat "nicer" looking, it's also much slower, and uses a lot more program space.
